CovidTransformer.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "80zQYJz7s0RC"
      },
      "source": [
        "## 1. Importing the Data\n",
        "Assuming the data files are stored in a local google drive storage, in a folder called \"data\"."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qA8yqaL-QuPf",
        "outputId": "f0c0d134-b172-4e1f-a208-bd42aee3b599"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /content/gdrive\n",
            "['id', 'tweet', 'label']\n",
            "['1', 'The CDC currently reports 99031 deaths. In general the discrepancies in death counts between different sources are small and explicable. The death toll stands at roughly 100000 people today.', 'real']\n",
            "['2', 'States reported 1121 deaths a small rise from last Tuesday. Southern states reported 640 of those deaths. https://t.co/YASGRTT4ux', 'real']\n",
            "['3', 'Politically Correct Woman (Almost) Uses Pandemic as Excuse Not to Reuse Plastic Bag https://t.co/thF8GuNFPe #coronavirus #nashville', 'fake']\n",
            "['4', '#IndiaFightsCorona: We have 1524 #COVID testing laboratories in India and as on 25th August 2020 36827520 tests have been done : @ProfBhargava DG @ICMRDELHI #StaySafe #IndiaWillWin https://t.co/Yh3ZxknnhZ', 'real']\n",
            "['5', 'Populous states can generate large case counts but if you look at the new cases per million today 9 smaller states are showing more cases per million than California or Texas: AL AR ID KS KY LA MS NV and SC. https://t.co/1pYW6cWRaS', 'real']\n"
          ]
        }
      ],
      "source": [
        "import csv\n",
        "import torch\n",
        "import torch.nn as nn\n",
        "import torch.nn.functional as F\n",
        "import torch.optim as optim\n",
        "import torchtext\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import math\n",
        "from torch.nn.utils.rnn import pad_sequence\n",
        "from google.colab import drive\n",
        "assert torch.cuda.is_available()\n",
        "\n",
        "drive.mount(\"/content/gdrive\")\n",
        "\n",
        "train_path = \"/content/gdrive/My Drive/data/Constraint_Train.csv\"\n",
        "valid_path = \"/content/gdrive/My Drive/data/Constraint_Val.csv\"\n",
        "test_path = \"/content/gdrive/My Drive/data/english_test_with_labels.csv\"\n",
        "\n",
        "def get_data(data_path):\n",
        "    # This is a very large file, so we will not load it into RAM\n",
        "    return csv.reader(open(data_path, \"rt\", encoding=\"utf-8\"))\n",
        "\n",
        "# Printing out some of the lines, to see if the CSV is being read correctly \n",
        "for i, line in enumerate(get_data(train_path)):\n",
        "    if i > 5:\n",
        "        break\n",
        "    print(line)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oB9uOKXxQuPi"
      },
      "source": [
        "The columns we care about are the second one and the last one. The second column are the social media posts, and the last column is the label indicating whether the post mentions \"real\" or \"fake\" news about Covid. Our task is to predict whether a social media post mentions real or fake news about Covid.\n",
        "\n",
        "We will need to split the text into words.\n",
        "We will do so by splitting at all whitespace\n",
        "characters.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "U5hkgqqLQuPi"
      },
      "outputs": [],
      "source": [
        "def split_post(post):\n",
        "    # separate punctuations\n",
        "    post = post.replace(\".\", \" . \") \\\n",
        "                 .replace(\",\", \" , \") \\\n",
        "                 .replace(\";\", \" ; \") \\\n",
        "                 .replace(\"?\", \" ? \")\n",
        "    return post.lower().split()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3fqAo8GhQuPk"
      },
      "source": [
        "## Extracting/Processing Data\n",
        "\n",
        "Since we are going to store the individual words in a post,\n",
        "we will defer looking up the word embeddings.\n",
        "Instead, we will store the **index** of each word in a PyTorch tensor, as it is\n",
        "more memory efficient."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ob36Cpz2QuPl",
        "outputId": "5e32638e-2cf4-4215-aaee-c872e12cb86b"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training dataset size: 5604\n",
            "Validation dataset size: 1873\n",
            "Test dataset size: 1848\n"
          ]
        }
      ],
      "source": [
        "def get_post_words(glove_vector, path):\n",
        "    idxs_li, labels_li = [], []\n",
        "    ignored_header = False\n",
        "    for line in get_data(path):\n",
        "      if not ignored_header:\n",
        "        ignored_header = True\n",
        "        continue   # Skipping header line in csv file\n",
        "      post = line[1]\n",
        "      if len(post) > 280:  # ignoring input posts longer than the 280 characters \n",
        "        continue            # (max size of a post), for efficiency\n",
        "                                      \n",
        "      idxs = [glove_vector.stoi[w]        # lookup the index of word\n",
        "               for w in split_post(post)\n",
        "               if w in glove_vector.stoi] # keep words that has an embedding\n",
        "      if not idxs: # ignore posts without any word with an embedding\n",
        "        continue                       \n",
        "      idxs = torch.tensor(idxs) # convert list to pytorch tensor\n",
        "      label = torch.tensor(int(line[2] == \"real\")).long()\n",
        "      idxs_li.append(idxs)\n",
        "      labels_li.append(label)\n",
        "    return idxs_li, labels_li\n",
        "\n",
        "# Adding padding to all the data sets\n",
        "idxs_train, labels_train = get_post_words(glove, train_path)\n",
        "idxs_train = pad_sequence(idxs_train, batch_first=True)  # Padding\n",
        "train_data = [[idxs_train[i], labels_train[i]] for i in range(len(labels_train))]\n",
        "\n",
        "idxs_valid, labels_valid = get_post_words(glove, valid_path)\n",
        "idxs_valid = pad_sequence(idxs_valid, batch_first=True) # Padding\n",
        "valid_data = [[idxs_valid[i], labels_valid[i]] for i in range(len(labels_valid))] \n",
        "\n",
        "idxs_test, labels_test = get_post_words(glove, test_path)\n",
        "idxs_test = pad_sequence(idxs_test, batch_first=True) # Padding\n",
        "test_data = [[idxs_test[i], labels_test[i]] for i in range(len(labels_test))] \n",
        "\n",
        "# Printing data set sizes after processing\n",
        "print(\"Training dataset size:\", len(train_data))\n",
        "print(\"Validation dataset size:\", len(valid_data))\n",
        "print(\"Test dataset size:\", len(test_data))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wd1a73-jQuPm"
      },
      "source": [
        "Here's what an element of the training set looks like:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "t8uVyL6vQuPn",
        "outputId": "f182369b-6a32-4cc6-d9dd-fabb1d251a72"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "tensor([     0,  10423,    839,    687,   1933,      2,      6,    216,      0,\n",
            "         24840,      6,    336,   4382,    118,    494,   1099,     32,    357,\n",
            "             5, 156629,      2,      0,    336,   2493,   2577,     22,   3208,\n",
            "        236671,     69,    373,      2,      0,      0,      0,      0,      0,\n",
            "             0,      0,      0,      0,      0,      0,      0,      0,      0,\n",
            "             0,      0,      0,      0,      0,      0,      0,      0,      0,\n",
            "             0,      0,      0,      0,      0,      0,      0,      0,      0,\n",
            "             0,      0,      0,      0,      0,      0])\n",
            "tensor(1)\n"
          ]
        }
      ],
      "source": [
        "post, label = train_data[0]\n",
        "print(post)\n",
        "print(label)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "26cprBTD_6Tp"
      },
      "source": [
        "# Building the Transformer Model\n",
        "Here we build the components of the Transformer Encoder, and in the end encapsulate the layers into a final Transformer Model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lxdxGE6XQuPp"
      },
      "source": [
        "## Input Embedding\n",
        "\n",
        "We are also going to use an `nn.Embedding` layer, instead of using the variable\n",
        "`glove` directly. The reason is that the `nn.Embedding` layer lets us look up\n",
        "the embeddings of **multiple words** simultaneously, so that our network\n",
        "can make predictions and train faster:"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "j-KLSxDsQuPq",
        "outputId": "18bc8fe0-66a7-4060-bea6-7eec0c2eedba"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "tensor([     0,  10423,    839,    687,   1933,      2,      6,    216,      0,\n",
            "         24840,      6,    336,   4382,    118,    494,   1099,     32,    357,\n",
            "             5, 156629,      2,      0,    336,   2493,   2577,     22,   3208,\n",
            "        236671,     69,    373,      2,      0,      0,      0,      0,      0,\n",
            "             0,      0,      0,      0,      0,      0,      0,      0,      0,\n",
            "             0,      0,      0,      0,      0,      0,      0,      0,      0,\n",
            "             0,      0,      0,      0,      0,      0,      0,      0,      0,\n",
            "             0,      0,      0,      0,      0,      0])\n",
            "tensor([[ 0.4180,  0.2497, -0.4124,  ..., -0.1841, -0.1151, -0.7858],\n",
            "        [ 1.0754, -0.8692,  0.5887,  ...,  0.9364,  0.4252,  1.4567],\n",
            "        [ 0.1419,  0.3410,  0.2213,  ...,  0.2460, -0.4449,  0.1761],\n",
            "        ...,\n",
            "        [ 0.4180,  0.2497, -0.4124,  ..., -0.1841, -0.1151, -0.7858],\n",
            "        [ 0.4180,  0.2497, -0.4124,  ..., -0.1841, -0.1151, -0.7858],\n",
            "        [ 0.4180,  0.2497, -0.4124,  ..., -0.1841, -0.1151, -0.7858]])\n"
          ]
        }
      ],
      "source": [
        "# Create an `nn.Embedding` layer and load data from pretrained `glove.vectors`\n",
        "glove_emb = nn.Embedding.from_pretrained(glove.vectors)\n",
        "print(post)\n",
        "# Example: we use the forward function of glove_emb to lookup the\n",
        "# embedding of each word in \"post\"\n",
        "post_emb = glove_emb(post)\n",
        "print(post_emb)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "aNRpnis_QuPu"
      },
      "source": [
        "This tensor is a representation of the entire post,\n",
        "and will be used as an input to our classifier.\n",
        "\n",
        "## Positional Encoding"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "faGsB1TfQuPu"
      },
      "outputs": [],
      "source": [
        "# Creating Positional Encoding Layer using code from https://pytorch.org/tutorials/beginner/transformer_tutorial.html\n",
        "class PositionalEncoding(nn.Module):\n",
        "  def __init__(self, emb_size, dropout = 0.1, max_len = 5000):\n",
        "    super(PositionalEncoding, self).__init__()\n",
        "    self.dropout = nn.Dropout(p=dropout)\n",
        "    position = torch.arange(max_len).unsqueeze(1)\n",
        "    div_term = torch.exp(torch.arange(0, emb_size, 2) * (-math.log(10000.0) / emb_size))\n",
        "    pe = torch.zeros(max_len, 1, emb_size)\n",
        "    pe[:, 0, 0::2] = torch.sin(position * div_term)\n",
        "    pe[:, 0, 1::2] = torch.cos(position * div_term)\n",
        "    self.register_buffer('pe', pe)\n",
        "\n",
        "  def forward(self, x):\n",
        "    x = x + self.pe[:x.size(0)]\n",
        "    x = self.dropout(x)\n",
        "    return x"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0gYm9YmfAH2n"
      },
      "source": [
        "##  Transformer Model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "I71vk6UKAKhK"
      },
      "outputs": [],
      "source": [
        "class Transformer(nn.Module):\n",
        "    def __init__(self, emb_size, dim_feedforward = 2048, num_layers = 2, dropout_pos = 0.1, dropout_encoder = 0.1):\n",
        "        super(Transformer, self).__init__()\n",
        "        self.emb_size = emb_size\n",
        "        self.emb = nn.Embedding.from_pretrained(glove.vectors, freeze=True)\n",
        "        self.pos_encode = PositionalEncoding(emb_size, dropout = dropout_pos)\n",
        "        self.encoder_layers = nn.TransformerEncoderLayer(emb_size, 5, dim_feedforward = dim_feedforward, \n",
        "                                                         batch_first=True, dropout = dropout_encoder)\n",
        "        self.encode = nn.TransformerEncoder(self.encoder_layers, num_layers)\n",
        "        self.linear = nn.Linear(emb_size, 2)\n",
        "    \n",
        "    def forward(self, x):\n",
        "        mask = (x == 0)\n",
        "        x = self.emb(x.cuda()) * math.sqrt(self.emb_size)\n",
        "        x = self.pos_encode(x.cuda())\n",
        "        x = self.encode(x.cuda(), src_key_padding_mask = mask)\n",
        "        x = x.mean(dim=1)\n",
        "        out = self.linear(x.cuda())\n",
        "        return out"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OXBEu63gAMWi"
      },
      "source": [
        "##  Training Functions"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vNkIdVZaQuP-"
      },
      "outputs": [],
      "source": [
        "def get_accuracy(model, data):\n",
        "    loader = torch.utils.data.DataLoader(data, batch_size=256)\n",
        "    correct, total = 0, 0\n",
        "    model.eval()\n",
        "    for posts, labels in loader:\n",
        "        posts = posts.cuda()\n",
        "        labels = labels.cuda()\n",
        "\n",
        "        output = model(posts)\n",
        "        pred = output.argmax(1)\n",
        "        correct += pred.eq(labels).sum().item()\n",
        "        total += labels.shape[0]\n",
        "    return correct / total"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "ghPIPBNqX138"
      },
      "outputs": [],
      "source": [
        "def train_network(model, train_data, valid_data, batch_size = 32, num_epochs=5, learning_rate=1e-3):\n",
        "    criterion = nn.CrossEntropyLoss()\n",
        "    optimizer = torch.optim.Adam(model.parameters(), lr=learning_rate)\n",
        "    losses, train_acc, valid_acc = [], [], []\n",
        "    epochs = []\n",
        "    train = torch.utils.data.DataLoader(train_data,\n",
        "                                               batch_size=batch_size,\n",
        "                                               shuffle=True)\n",
        "    valid = torch.utils.data.DataLoader(valid_data,\n",
        "                                               batch_size=batch_size,\n",
        "                                               shuffle=True)\n",
        "    \n",
        "    checkpoint_path = '/content/gdrive/My Drive/checkpointing/ckpt-{}.pk'\n",
        "    \n",
        "    model = model.cuda()\n",
        "    model.train()\n",
        "    # Training\n",
        "    for epoch in range(num_epochs):\n",
        "        for posts, labels in train:\n",
        "            posts = posts.cuda()\n",
        "            labels = labels.cuda()\n",
        "\n",
        "            optimizer.zero_grad()\n",
        "            pred = model(posts)\n",
        "            loss = criterion(pred, labels)\n",
        "            loss.backward()\n",
        "            optimizer.step()\n",
        "        losses.append(float(loss))\n",
        "\n",
        "        epochs.append(epoch)\n",
        "        train_acc.append(get_accuracy(model, train_data))\n",
        "        valid_acc.append(get_accuracy(model, valid_data))\n",
        "        print(\"Epoch %d; Loss %f; Train Acc %f; Val Acc %f\" % (\n",
        "              epoch+1, loss, train_acc[-1], valid_acc[-1]))\n",
        "        \n",
        "        # Checkpointing\n",
        "        # print(\"Saving Model........\")\n",
        "        torch.save(model.state_dict(), checkpoint_path.format(epoch)) \n",
        "\n",
        "    # Plotting\n",
        "    plt.title(\"Training Curve\")\n",
        "    plt.plot(losses, label=\"Train\")\n",
        "    plt.xlabel(\"Epoch\")\n",
        "    plt.ylabel(\"Loss\")\n",
        "    plt.show()\n",
        "\n",
        "    plt.title(\"Training Curve\")\n",
        "    plt.plot(epochs, train_acc, label=\"Train\")\n",
        "    plt.plot(epochs, valid_acc, label=\"Validation\")\n",
        "    plt.xlabel(\"Epoch\")\n",
        "    plt.ylabel(\"Accuracy\")\n",
        "    plt.legend(loc='best')\n",
        "    plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zuj5ELdGujik"
      },
      "source": [
        "## Overfit to a Small Dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 920
        },
        "id": "l6MGFMnJulVc",
        "outputId": "322b5b6a-f881-42db-e49d-746c1162f189"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1; Loss 0.593686; Train Acc 0.720000; Val Acc 0.720000\n",
            "Epoch 2; Loss 1.058269; Train Acc 0.720000; Val Acc 0.720000\n",
            "Epoch 3; Loss 0.636090; Train Acc 0.760000; Val Acc 0.760000\n",
            "Epoch 4; Loss 0.364439; Train Acc 0.940000; Val Acc 0.940000\n",
            "Epoch 5; Loss 0.251988; Train Acc 0.940000; Val Acc 0.940000\n",
            "Epoch 6; Loss 0.254728; Train Acc 0.980000; Val Acc 0.980000\n",
            "Epoch 7; Loss 0.176244; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 8; Loss 0.048285; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 9; Loss 0.030777; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 10; Loss 0.038201; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 11; Loss 0.012006; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 12; Loss 0.004983; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 13; Loss 0.005259; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 14; Loss 0.003409; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 15; Loss 0.002709; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 16; Loss 0.002324; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 17; Loss 0.001895; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 18; Loss 0.001779; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 19; Loss 0.001743; Train Acc 1.000000; Val Acc 1.000000\n",
            "Epoch 20; Loss 0.001427; Train Acc 1.000000; Val Acc 1.000000\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "# Initial testing to see if model/training works\n",
        "train_data_subset = torch.utils.data.Subset(train_data, np.random.choice(len(train_data), 50, replace=False))\n",
        "model = Transformer(50, 2).cuda()\n",
        "train_network(model, train_data_subset, train_data_subset, batch_size = 20, num_epochs=20, learning_rate=0.003)\n",
        "# Note here the validation and training accuracy will be the same since the same dataset is passsed in for both parameters "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IH4CRakbWk6m"
      },
      "source": [
        "## Test if Working on Entire Dataset"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "Xuz5VWFeQuQB",
        "outputId": "f7aef335-56bf-4404-9e35-0b324acaf3c4"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1; Loss 0.618732; Train Acc 0.725018; Val Acc 0.735184\n",
            "Epoch 2; Loss 0.481036; Train Acc 0.802463; Val Acc 0.793914\n",
            "Epoch 3; Loss 0.359928; Train Acc 0.821199; Val Acc 0.826482\n",
            "Epoch 4; Loss 0.395581; Train Acc 0.849036; Val Acc 0.834490\n",
            "Epoch 5; Loss 0.356034; Train Acc 0.853854; Val Acc 0.853711\n",
            "Epoch 6; Loss 0.325188; Train Acc 0.869022; Val Acc 0.855312\n",
            "Epoch 7; Loss 0.303962; Train Acc 0.873840; Val Acc 0.866524\n",
            "Epoch 8; Loss 0.244219; Train Acc 0.883833; Val Acc 0.872931\n",
            "Epoch 9; Loss 0.239191; Train Acc 0.888829; Val Acc 0.871863\n",
            "Epoch 10; Loss 0.251982; Train Acc 0.890971; Val Acc 0.884677\n",
            "Epoch 11; Loss 0.255554; Train Acc 0.894897; Val Acc 0.873999\n",
            "Epoch 12; Loss 0.228943; Train Acc 0.905782; Val Acc 0.893753\n",
            "Epoch 13; Loss 0.253564; Train Acc 0.911849; Val Acc 0.894821\n",
            "Epoch 14; Loss 0.216516; Train Acc 0.917380; Val Acc 0.890016\n",
            "Epoch 15; Loss 0.232046; Train Acc 0.920414; Val Acc 0.899092\n",
            "Epoch 16; Loss 0.208921; Train Acc 0.908994; Val Acc 0.895355\n",
            "Epoch 17; Loss 0.178271; Train Acc 0.924161; Val Acc 0.901762\n",
            "Epoch 18; Loss 0.187964; Train Acc 0.929872; Val Acc 0.900694\n",
            "Epoch 19; Loss 0.193959; Train Acc 0.936652; Val Acc 0.899626\n",
            "Epoch 20; Loss 0.169750; Train Acc 0.939507; Val Acc 0.906567\n",
            "Epoch 21; Loss 0.161279; Train Acc 0.943969; Val Acc 0.907635\n",
            "Epoch 22; Loss 0.137449; Train Acc 0.942541; Val Acc 0.904965\n",
            "Epoch 23; Loss 0.113285; Train Acc 0.953605; Val Acc 0.909237\n",
            "Epoch 24; Loss 0.166492; Train Acc 0.951106; Val Acc 0.909770\n",
            "Epoch 25; Loss 0.150799; Train Acc 0.952355; Val Acc 0.906567\n",
            "Epoch 26; Loss 0.142726; Train Acc 0.961456; Val Acc 0.910838\n",
            "Epoch 27; Loss 0.102012; Train Acc 0.949679; Val Acc 0.906033\n",
            "Epoch 28; Loss 0.110418; Train Acc 0.961991; Val Acc 0.911372\n",
            "Epoch 29; Loss 0.124366; Train Acc 0.961278; Val Acc 0.905499\n",
            "Epoch 30; Loss 0.101310; Train Acc 0.967880; Val Acc 0.908703\n",
            "Epoch 31; Loss 0.111127; Train Acc 0.973769; Val Acc 0.913508\n",
            "Epoch 32; Loss 0.071952; Train Acc 0.977516; Val Acc 0.914042\n",
            "Epoch 33; Loss 0.085984; Train Acc 0.977159; Val Acc 0.915643\n",
            "Epoch 34; Loss 0.060004; Train Acc 0.980728; Val Acc 0.909237\n",
            "Epoch 35; Loss 0.033630; Train Acc 0.984297; Val Acc 0.914042\n",
            "Epoch 36; Loss 0.067473; Train Acc 0.985724; Val Acc 0.914576\n",
            "Epoch 37; Loss 0.082956; Train Acc 0.985011; Val Acc 0.912440\n",
            "Epoch 38; Loss 0.070764; Train Acc 0.988044; Val Acc 0.916711\n",
            "Epoch 39; Loss 0.029621; Train Acc 0.986438; Val Acc 0.914576\n",
            "Epoch 40; Loss 0.072278; Train Acc 0.987152; Val Acc 0.907101\n",
            "Epoch 41; Loss 0.031984; Train Acc 0.990007; Val Acc 0.917245\n",
            "Epoch 42; Loss 0.032615; Train Acc 0.989650; Val Acc 0.906567\n",
            "Epoch 43; Loss 0.066819; Train Acc 0.994825; Val Acc 0.911372\n",
            "Epoch 44; Loss 0.019856; Train Acc 0.993933; Val Acc 0.911906\n",
            "Epoch 45; Loss 0.042116; Train Acc 0.995182; Val Acc 0.912974\n",
            "Epoch 46; Loss 0.025082; Train Acc 0.993219; Val Acc 0.910838\n",
            "Epoch 47; Loss 0.014319; Train Acc 0.996074; Val Acc 0.911372\n",
            "Epoch 48; Loss 0.009988; Train Acc 0.995896; Val Acc 0.913508\n",
            "Epoch 49; Loss 0.020258; Train Acc 0.995896; Val Acc 0.910304\n",
            "Epoch 50; Loss 0.046213; Train Acc 0.996431; Val Acc 0.910304\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "model = Transformer(50).cuda()\n",
        "train_network(model, train_data, valid_data, batch_size = 512, num_epochs=50, learning_rate=2e-4)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E97RislODN38"
      },
      "source": [
        "# Hyperparameter Tuning"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = Transformer(50).cuda()\n",
        "train_network(model, train_data, valid_data, batch_size = 512, num_epochs=50, learning_rate=1e-4)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "DLAa-IWIFJTa",
        "outputId": "069310ca-4d4b-45b2-bf72-08b9735c91cc"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1; Loss 0.662430; Train Acc 0.641328; Val Acc 0.669514\n",
            "Epoch 2; Loss 0.567153; Train Acc 0.741435; Val Acc 0.752803\n",
            "Epoch 3; Loss 0.495661; Train Acc 0.766774; Val Acc 0.772557\n",
            "Epoch 4; Loss 0.442044; Train Acc 0.799251; Val Acc 0.799786\n",
            "Epoch 5; Loss 0.444437; Train Acc 0.817452; Val Acc 0.815804\n",
            "Epoch 6; Loss 0.380095; Train Acc 0.835832; Val Acc 0.834490\n",
            "Epoch 7; Loss 0.322434; Train Acc 0.843862; Val Acc 0.844100\n",
            "Epoch 8; Loss 0.363260; Train Acc 0.852784; Val Acc 0.856380\n",
            "Epoch 9; Loss 0.369576; Train Acc 0.861349; Val Acc 0.860117\n",
            "Epoch 10; Loss 0.308193; Train Acc 0.867773; Val Acc 0.866524\n",
            "Epoch 11; Loss 0.269740; Train Acc 0.875446; Val Acc 0.868126\n",
            "Epoch 12; Loss 0.294537; Train Acc 0.880799; Val Acc 0.870796\n",
            "Epoch 13; Loss 0.242080; Train Acc 0.884547; Val Acc 0.868126\n",
            "Epoch 14; Loss 0.246434; Train Acc 0.889186; Val Acc 0.871863\n",
            "Epoch 15; Loss 0.299608; Train Acc 0.894540; Val Acc 0.878270\n",
            "Epoch 16; Loss 0.221086; Train Acc 0.896502; Val Acc 0.872931\n",
            "Epoch 17; Loss 0.243924; Train Acc 0.893469; Val Acc 0.880406\n",
            "Epoch 18; Loss 0.185592; Train Acc 0.905425; Val Acc 0.882007\n",
            "Epoch 19; Loss 0.217718; Train Acc 0.910600; Val Acc 0.884143\n",
            "Epoch 20; Loss 0.215585; Train Acc 0.910600; Val Acc 0.887880\n",
            "Epoch 21; Loss 0.197354; Train Acc 0.905960; Val Acc 0.874533\n",
            "Epoch 22; Loss 0.208487; Train Acc 0.917024; Val Acc 0.883609\n",
            "Epoch 23; Loss 0.200233; Train Acc 0.912206; Val Acc 0.880406\n",
            "Epoch 24; Loss 0.253320; Train Acc 0.925946; Val Acc 0.886813\n",
            "Epoch 25; Loss 0.230303; Train Acc 0.918094; Val Acc 0.878804\n",
            "Epoch 26; Loss 0.174607; Train Acc 0.920771; Val Acc 0.887880\n",
            "Epoch 27; Loss 0.230916; Train Acc 0.933262; Val Acc 0.890550\n",
            "Epoch 28; Loss 0.222913; Train Acc 0.931834; Val Acc 0.882541\n",
            "Epoch 29; Loss 0.178796; Train Acc 0.922020; Val Acc 0.886279\n",
            "Epoch 30; Loss 0.183701; Train Acc 0.938615; Val Acc 0.885211\n",
            "Epoch 31; Loss 0.149228; Train Acc 0.941292; Val Acc 0.886279\n",
            "Epoch 32; Loss 0.161476; Train Acc 0.946645; Val Acc 0.891084\n",
            "Epoch 33; Loss 0.140606; Train Acc 0.947002; Val Acc 0.890550\n",
            "Epoch 34; Loss 0.147300; Train Acc 0.951820; Val Acc 0.892152\n",
            "Epoch 35; Loss 0.152737; Train Acc 0.948251; Val Acc 0.890550\n",
            "Epoch 36; Loss 0.104423; Train Acc 0.949143; Val Acc 0.888948\n",
            "Epoch 37; Loss 0.137691; Train Acc 0.952891; Val Acc 0.887880\n",
            "Epoch 38; Loss 0.109269; Train Acc 0.955211; Val Acc 0.888414\n",
            "Epoch 39; Loss 0.136632; Train Acc 0.952891; Val Acc 0.888414\n",
            "Epoch 40; Loss 0.113390; Train Acc 0.951642; Val Acc 0.887347\n",
            "Epoch 41; Loss 0.107348; Train Acc 0.966631; Val Acc 0.892686\n",
            "Epoch 42; Loss 0.104856; Train Acc 0.965560; Val Acc 0.892686\n",
            "Epoch 43; Loss 0.083405; Train Acc 0.968059; Val Acc 0.889482\n",
            "Epoch 44; Loss 0.084942; Train Acc 0.967702; Val Acc 0.892686\n",
            "Epoch 45; Loss 0.091590; Train Acc 0.965739; Val Acc 0.889482\n",
            "Epoch 46; Loss 0.103755; Train Acc 0.960029; Val Acc 0.883609\n",
            "Epoch 47; Loss 0.086174; Train Acc 0.975196; Val Acc 0.888948\n",
            "Epoch 48; Loss 0.069659; Train Acc 0.975732; Val Acc 0.891618\n",
            "Epoch 49; Loss 0.090056; Train Acc 0.977873; Val Acc 0.888414\n",
            "Epoch 50; Loss 0.065831; Train Acc 0.978051; Val Acc 0.891618\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXxU5fX48c/JHkhCyMoSIFF2ZRMEFBfccQN3wQ3USrXV1qptbX9qFbW1ra3W9SuuuCKKIiouiFKwghJ2CKCsIWEPWYEskzm/P+4FxjCEAJlMMjnv1yuvzN1mzg3DPfdZ7vOIqmKMMcbUFBbsAIwxxjROliCMMcb4ZQnCGGOMX5YgjDHG+GUJwhhjjF+WIIwxxvhlCcI0eyLymYiMru99jWnqxJ6DME2RiJT5LLYAKoBqd/mXqvpWw0d1dEQkARgHXAYkAVuBj4FHVHVHMGMzzZOVIEyTpKpxe3+AXOBin3X7koOIRAQvyroTkShgBnAcMAxIAE4CCoCBR/B+TeK8TeNmCcKEFBEZKiJ5IvJHEdkCvCoirUXkExHZLiKF7usMn2Nmisgv3NdjRORbEXnc3XediJx/hPtmicgsESkVka9E5FkRefMgod8AdAQuVdUcVfWq6jZVfVhVp7nvpyLS2ef9XxORR2o57xUicpHP/hHu3+AEd3mwiHwnIkUislhEhh7t39+EFksQJhS1wami6QSMxfmev+oudwT2AM/UcvwgYBWQAvwDeFlE5Aj2fRv4AUgGHgSur+UzzwY+V9WyWvY5lJrn/Q4wymf7ecAOVV0gIu2BT4FH3GPuASaLSOpRfL4JMZYgTCjyAn9R1QpV3aOqBao6WVV3q2op8Chwei3Hb1DVF1W1GpgAtAXSD2dfEekInAg8oKqVqvotMLWWz0wGNh/eaR7gZ+eNk6CGi0gLd/s1OEkD4DpgmqpOc0sr04Fs4IKjjMGEEEsQJhRtV9XyvQsi0kJEXhCRDSJSAswCEkUk/CDHb9n7QlV3uy/jDnPfdsBOn3UAG2uJuQAnuRyNn523qq4GVgAXu0liOE7SAKeUcaVbvVQkIkXAKfUQgwkh1pBlQlHNrnl3A92AQaq6RUT6AguBg1Ub1YfNQJKItPBJEh1q2f8r4BERaamquw6yz26cHlt7tQHyfJb9dUncW80UBuS4SQOcZPWGqt5yiPMwzZiVIExzEI/T7lAkIknAXwL9gaq6AafK5kERiRKRk4CLaznkDZyL9mQR6S4iYSKSLCJ/FpG91T6LgGtEJFxEhlF7NdleE4FzgdvYX3oAeBOnZHGe+34xbkN3ht93Mc2SJQjTHDwJxAI7gLnA5w30udeyv6vqI8C7OM9rHEBVK3AaqlcC04ESnAbuFOB7d7ff4iSZIve9pxwqAFXdDMwBTnY/f+/6jcAI4M/Adpzk9HvsmmB82INyxjQQEXkXWKmqAS/BGFMf7G7BmAARkRNF5Fi3umgYzh37Ie/6jWksrJHamMBpA3yA04U1D7hNVRcGNyRj6s6qmIwxxvhlVUzGGGP8CpkqppSUFM3MzAx2GMYY06TMnz9/h6r6HWIlZBJEZmYm2dnZwQ7DGGOaFBHZcLBtVsVkjDHGL0sQxhhj/LIEYYwxxq+QaYPwp6qqiry8PMrLyw+9s6mTmJgYMjIyiIyMDHYoxpgAC+kEkZeXR3x8PJmZmRx8vhdTV6pKQUEBeXl5ZGVlBTscY0yAhXQVU3l5OcnJyZYc6omIkJycbCUyY5qJkE4QgCWHemZ/T2Oaj5CuYjLGmKbO61XWFexiWX4xxXuq8FQr1V6lyuululrxeJX0hBiuGdSx3j/bEkQAFRQUcNZZZwGwZcsWwsPDSU11Hlj84YcfiIqKOuix2dnZvP766zz11FMNEqsxJnhUlT1V1ZRVeCjZU0XO5lKW5RezJK+IZfkllFV4aj2+X8dESxBNTXJyMosWLQLgwQcfJC4ujnvuuWffdo/HQ0SE/3+CAQMGMGDAgAaJ0xjTsBbkFnL/lGUUlFWyq8LDrkoP3hrjpkZFhNGjbQKX9mtPr4xW9GrfitT4aCLChIjwMCLChPAwISJMAlb1awmigY0ZM4aYmBgWLlzIkCFDGDlyJL/97W8pLy8nNjaWV199lW7dujFz5kwef/xxPvnkEx588EFyc3NZu3Ytubm53HnnnfzmN78J9qkYY45AaXkVd7y9EK8qp3VNoWV0BHHREbSMjqBlVDjxMZF0SY+ja3o8keHBbSZuNgnioY+Xk7OppF7fs2e7BP5y8XGHfVxeXh7fffcd4eHhlJSUMHv2bCIiIvjqq6/485//zOTJkw84ZuXKlXzzzTeUlpbSrVs3brvtNnsWwZgm6KGPc9hcvIf3bj2Z/p1aBzucWjWbBNGYXHnllYSHhwNQXFzM6NGj+emnnxARqqqq/B5z4YUXEh0dTXR0NGlpaWzdupWMDJtf3pim5PNlm3l/fh6/ObNzo08O0IwSxJHc6QdKy5Yt972+//77OeOMM/jwww9Zv349Q4cO9XtMdHT0vtfh4eF4PLU3WhljDk+1V/lxaynz1u9kWX4xADGR4cREhhMdEbbvd8ekFvRom0BG69jDqvvfVlLOnz5YSu+MVtxxVpdAnUa9ajYJorEqLi6mffv2ALz22mvBDcaYEPRe9kbey84jJT6KtPgY0hKiSY+PIT0hhvAwYUFuIdnrd5K9oZDScufGK7llFJHhYZR7qqmo8lLuqabm5JsJMRF0b5tAT/fnlC4ptEuM9RuDqvKHyUvYU1XNE1f3DXrbQl1ZggiyP/zhD4wePZpHHnmECy+8MNjhGBNSluQV8acPlpLROpaCXRXM/nEHpX66jHZJi+Oi3u04MbM1J2YmHVA6UFUqq73sqaxm7Y5drNhcQs6mElZsLmFS9kZ2V1YTFR7GdYM7cfuZnUlq+fMu7G/O3cDMVdsZN+I4jk2NC/h515eQmZN6wIABWnPCoBUrVtCjR48gRRS67O9qmoLdlR4ueupb9lRV8/lvT6NVC6dTx64KD9tKK9haUk55VTW9MxIPuKAfDq9XWbujjJdmr2NS9kZaRkXwy9OP4aZTsmgRFcGa7WVc+NRsBmYlM+HGExvdaAQiMl9V/faptxKEMSYkPfLpCtYV7OKtXwzalxwAWkZHkBUdQVZKy1qOrruwMKFzWjyPXd6bX5yaxd8/X8XjX/7I63M28NuzuzBp3kZiIsP55xW9G11yOBRLEMaYkDM9Zytvf5/LL087hpOPTWmwz+2cFs+LNwwge/1O/vbZSv7fh8sAeP7aE0hPiGmwOOqLJQhjTEjZVlrOHycvoWfbBO46t2tQYhiQmcT7t57EjBXb2F5Wwfm92gYljqNlCcIY0+ioKqUVHuKjIw6rWkZV+cP7S9hV4eE/I/sSHREewChrJyKc3TM9aJ9fHyxBGGOCbnelh6V5xSzILWJhbiELNxaxvbSCtPho+nVMpF/H1pzQsTW92rciNurgF/033N5CDw0/ji7p8Q14BqEpoAlCRIYB/wHCgZdU9bEa2zsBrwCpwE7gOlXNc7dVA0vdXXNVdXggYzXG1N2eympiIsOOqtF1zfYyPlqYz4yV21i5pZRqd7S6zOQWnNI5hc5pcfy0tZSFG4v4YvlWAMLDhO5t4jk2NY52ibG0T4yhXWIs7RJjqfR4efTTFZzeNZUbTupUL+fZ3AUsQYhIOPAscA6QB8wTkamqmuOz2+PA66o6QUTOBP4GXO9u26OqfQMVX0M544wzuPfeeznvvPP2rXvyySdZtWoVzz///AH7Dx06lMcff5wBAwZwwQUX8Pbbb5OYmPizffyNDFvTlClT6Nq1Kz179gTggQce4LTTTuPss8+upzMzzdHijUWMn72Wz5ZupkNSCy7s1ZYLe7elZ9uEOiWLrSXlfLx4Ex8t2sTS/GJE4MTMJG47/Vj6dUykb4dEkuOiDziuoKyCRRuLWJhbxOI85+ezZZupqv55N/2kllH888qm11uosQpkCWIgsFpV1wKIyERgBOCbIHoCd7mvvwGmBDCeoBg1ahQTJ078WYKYOHEi//jHPw557LRp0474c6dMmcJFF120L0GMGzfuiN/LNG9er/LNqm28MGstP6zbSXx0BNcP7sTaHbt4YdZanpu5hqyUlvuSRZuEGAp2VbJzVyUFZRUU7KqkoKySH9YX8N2aAlShd0Yr7ruwB8P7tCOtDr17kuOiOatHOmf12F+n7/UqO8oqyC/aw6aicjYX7+HULqmkxTe93kKNVSATRHtgo89yHjCoxj6LgctwqqEuBeJFJFlVC4AYEckGPMBjqnpA8hCRscBYgI4d63+yjPpwxRVXcN9991FZWUlUVBTr169n06ZNvPPOO9x1113s2bOHK664goceeuiAYzMzM8nOziYlJYVHH32UCRMmkJaWRocOHejfvz8AL774IuPHj6eyspLOnTvzxhtvsGjRIqZOncp///tfHnnkESZPnszDDz/MRRddxBVXXMGMGTO455578Hg8nHjiiTz//PNER0eTmZnJ6NGj+fjjj6mqquK9996je/fuDf0nMw1MVSncXcWeqmrKq/YPLVFR5WXdjl288r91rN5WRrtWMdx3YQ+uPrED8THOcwU7d1Xy+bItfLp0E8/NXM0z36w+6OdkpbTkjjO7MKJvu3p5mjgsTEhLiCEtIYZ+jfO/f5MX7Ebqe4BnRGQMMAvIB6rdbZ1UNV9EjgG+FpGlqrrG92BVHQ+MB+dJ6lo/6bN7YcvSWnc5bG16wfmP1bpLUlISAwcO5LPPPmPEiBFMnDiRq666ij//+c8kJSVRXV3NWWedxZIlS+jdu7ff95g/fz4TJ05k0aJFeDweTjjhhH0J4rLLLuOWW24B4L777uPll1/mjjvuYPjw4fsSgq/y8nLGjBnDjBkz6Nq1KzfccAPPP/88d955JwApKSksWLCA5557jscff5yXXnrpaP9KppFSVWau2s4TX/3Ikrzig+7Xs20C/xnZlwt6tT1gDKGkllFcM6gj1wzqyI6yCqbnbGVPZTXJcVEkt4x2f0fR2h3byDQtgUwQ+UAHn+UMd90+qroJpwSBiMQBl6tqkbst3/29VkRmAv2AnyWIpmJvNdPeBPHyyy8zadIkxo8fj8fjYfPmzeTk5Bw0QcyePZtLL72UFi1aADB8+P72+mXLlnHfffdRVFREWVnZz6qy/Fm1ahVZWVl07er0Dx89ejTPPvvsvgRx2WWXAdC/f38++OCDoz530/ioKrN+2sET039k0cYiMlrH8sdh3UlqGUl0RDgxkWFER4YTExFOq9hIerSNr1OdfkpcNKMG2q18KAlkgpgHdBGRLJzEMBK4xncHEUkBdqqqF/gTTo8mRKQ1sFtVK9x9hgCHrrSvzSHu9ANpxIgR/O53v2PBggXs3r2bpKQkHn/8cebNm0fr1q0ZM2YM5eXlR/TeY8aMYcqUKfTp04fXXnuNmTNnHlWse4cVtyHFQ4+q8r/VBfx7+ioW5BbRPjGWxy7rxeX9M+zu3vgVsG+FqnqA24EvgBXAJFVdLiLjRGTvLfBQYJWI/AikA4+663sA2SKyGKfx+rEavZ+alLi4OM444wxuuukmRo0aRUlJCS1btqRVq1Zs3bqVzz77rNbjTzvtNKZMmcKePXsoLS3l448/3rettLSUtm3bUlVVxVtvvbVvfXx8PKWlpQe8V7du3Vi/fj2rVzt1xW+88Qann356PZ2paWwK3TaCB6cu59wnZnHdy9+zubicRy89nm/uGcrIgR0tOZiDCmgbhKpOA6bVWPeAz+v3gff9HPcd0CuQsTW0UaNGcemllzJx4kS6d+9Ov3796N69Ox06dGDIkCG1HnvCCSdw9dVX06dPH9LS0jjxxBP3bXv44YcZNGgQqampDBo0aF9SGDlyJLfccgtPPfUU77+//08cExPDq6++ypVXXrmvkfrWW28NzEmboJi/oZBPl2xmztoCVm4pQRViI8MZkNmaG07O5KoBGUF9wtg0HTbctzls9ndtvFZsLuHip78lPEwYkNmawVnJnHRsMr0zEomKsJKCOZAN921MM1DtVe6dvIRWsZFMv+v0o5rjwBgIYBuEMabuPNVeVm8rZZef2c7qasJ361mcV8wDF/e05GDqRciXIFTVHruvR6FSJdmYVHique6l75m3vhBwuot2TIqlU3JLOiS1oGfbeM47rk2t3+O8wt08/uUqhnZLZXifdg0VuglxIZ0gYmJiKCgoIDk52ZJEPVBVCgoKiImxoQzqi6py7+SlzFtfyF3ndCU8TMgt2E3uzt38sG4nUxblowpXD+jAXy/rRXjYgd9jVeX+Kc7ENI9ccrx91029CekEkZGRQV5eHtu3bw92KCEjJiaGjIyMYIcRMp75ejUfLsznrnO68puzuhywvdLj5emvf+Lpr1dTUl7Fk37mOJi6eBPfrNrOAxf1JKN1i4YK3TQDIZ0gIiMjycrKCnYYxvj1yZJN/Gv6j1zarz13nNnZ7z5REWHcfW43EltE8fAnOZS+ls0L1/enZbTzX7dwVyXjPs6hT4dERp+c2YDRm+bAGqmNOULL8osZ9uQs3p2Xe9htMwtzC7l70mIGdGrNY5f3OmS10M2nZPH4lX2Ys7aAa1/6nqLdlQA8Om0FxXuqeOwg1U/GHA1LEMYcgeLdVdz65nx+2lbGHycv5c53F1FWxx5IeYW7ueX1+aQnxPDC9f3r/NDaFf0zeP7aE8jZXMJVL8zhgwV5vD8/j7GnHUOPtglHczrG+GUJwpjD5PUqd01axNaScib9cjB3n9OVjxdv4uKnv2VZ/sFHRQUoq/DwiwnZVHiqeWXMAL+T49Tm3OPa8NqNJ5JfuIe7Ji0mK6Wl37YLY+qDJQhjDtPz/13DjJXbuO/CnvTvlMQdZ3XhnVsGs7vSw2XPfcfrc9b/rMqprMLDN6u28bdpKxjxzLf8tK2M5649gc5pRzZn8snHpvDO2MGc0DGRf17Rm5hIGzbDBEZID7VhTH37bs0Ornvpey7s3Y6nRvb9WdtBQVkFd7+3mJmrtnPecekckxrHnDUFLM0vptqrRIYL/Tq0ZsyQTC7o1TaIZ2HMfrUNtWEJwhjX1MWbWLKxiOtP6kSn5JYHbN9SXM5FT88msUUUH/16yL6eRL68XuWlb9fyj89XAdCnQyInHZPM4GOS6d+pNbFRdrdvGhcbi8mYQ5iyMJ/fTVqEKrzyv3UMO74Nt5x6DP06tgagqtrL7W8vYHdlNRPHnuA3OYAzDebY047l8hMyiIkMP+h+xjQF9u01zd7ny7Zw93uLGZSVxD8u78M783J5c+4Gpi3dwsDMJMaedgzfrSkge0MhT43qV6e2g8NtfDamMbIqJtOsfbNqG2Nfz+b49q144+ZBxLl3/GUVHt6dt5FXvl1HftEeAEaf1ImHRhwfzHCNqXdWxWRC0s5dlbz87Vo+XJBPZbX3gO0iwmldUrnplEyOa9fqgO1z1hRw6xvz6Zoez2s3DtyXHADioiO4+ZQsRp/UiU+XbmbF5lLuOqdrQM/HmMbGShCmydleWsGLs9fy5twN7Kmq5qzuaaQnHDiA4K4KD1/mbGV3ZTWDj0nipiFZnNUjnfAwYUFuIde99D3tE2N595cn2fDYptmyEoQJCVuKy3lh1hre/j6Xqmovw/u04/YzO9faJlC8u4p3s3OZ8N0Gxr4xn07JLbj8hAxenL2WtPho3vrFIEsOxhyElSBMk/DuvFzun7KcalUu69eeX53RmayUA7uiHoyn2suXOVt55dt1ZG8opH1iLJNuPYn2ibEBjNqYxi9oJQgRGQb8BwgHXlLVx2ps7wS8AqQCO4HrVDXP3TYauM/d9RFVnRDIWE3jNT1nK3/6YClDOqfw10t70SHp8Ie0jggP44JebbmgV1tyNpWQEh9FWrzNa2FMbQKWIEQkHHgWOAfIA+aJyFRVzfHZ7XHgdVWdICJnAn8DrheRJOAvwABAgfnusYWBitc0TgtyC7njnQX0at+KF67vT4uoo//K9mxnA9sZUxeBHItpILBaVdeqaiUwERhRY5+ewNfu6298tp8HTFfVnW5SmA4MC2CsphFat2MXv5iQTXpCDC+PObFekoMxpu4CmSDaAxt9lvPcdb4WA5e5ry8F4kUkuY7HIiJjRSRbRLJt1rjQsqOsgjGv/gDAhBsHkmIPnhnT4IJ9S3YP8IyIjAFmAflAdV0PVtXxwHhwGqkDEaCpP3sqq1m5pYSczSX8uKWUtIQYTu+aynHtEn426N3uSg83vzaPrSXlvHPLYDIPozHaGFN/Apkg8oEOPssZ7rp9VHUTbglCROKAy1W1SETygaE1jp0ZwFhNABTvruK9+RtZuLGIFZtLWLdjF3s7zbWMCmdXZTX//GIVqfHRDO2aytBuaZx8bDJ3v7eYpfnFvHD9gH1jIRljGl4gE8Q8oIuIZOEkhpHANb47iEgKsFNVvcCfcHo0AXwB/FVE9l4dznW3myagpLyKV75dx8vfrqO03ENG61h6tk3g4t7t6NkugZ5tE8hoHcv2sgpm/biDb1Zt44vlW3hvft6+93jkkuM5p2d6EM/C1KqiDCJbQFgTmFJmaw7k/eB/W1QcZJ0OcamHfp89hbD+W0hoB237NY1zP0oBSxCq6hGR23Eu9uHAK6q6XETGAdmqOhWnlPA3EVGcKqZfu8fuFJGHcZIMwDhV3RmoWE39KC2v4rX/refF2WspKfdw3nHp/PasrgftNZQWH8MV/TO4on8GnmovizYWMXPVdtJbxXDd4E4NHH2QVZUDCpFBeC5j00LYMAeiWjgXzOh493cchEXAznVQsBoKfoId7u/dBU6CSD4WkjtDchdI6eK+7gwxjaCn2NYc+O9jkPPRIXYUyDgRug2DbhdAanfYW+VZsAZ+/BxWfQYbvgN1a8Dj2kDX85z9jzn96P/dqvbAxu+dBLRpIaQfD93Od+IKO8gQ8eXFsHqGE1tkLAx/6uhi8MMelDNHraS8ijfmbODF2Wsp2l3F2T3SufPsLhzf/sDxj4zLUwF52c4FYf1s2PgDRETDiGegZ83OfgHg9cLq6fDd087n10XLtP1JoHUn2LUDdvzkJI+iDaA+42HFpbtJw00eyZ0hKQvi20B0wv4LcCBsWwn//Tss/9BJdINvhb7XOn/fmsq2wU9fwqppzoUZILETdBoC+fNhhzOvB2k9oesw6HIOFOU6+6/+GipLISIWjj0D+oyCHhfX7dy81ZA7B9bNcr4DefOguhIk3PkbF6wGrwdaJDuf23UYHHsm7N4Bqz6HHz9zjtu7z/GXwwX/PKI/l00YZAJi/Y5dvPbdet7L3siuymrO7J7GnWd3oXdGYnADU4XNi+DHL6DXlc5dbmOx6B1Y/LaTEDzlgECbXpB1mnPByJ8Pg26Dc8ZBRACGAPFUwJJ34btnnItfQgYMvg16XeFcbCrKoHKXc+GrKHMuWq07uaWCWhK+p8ItabgJY29JY8dPsKdG4T+yhZNA4ts6CSOmlfuZPp9bWQaVu3Eeg/IjNsk5Nr6N+z7p0DIVcqbCsskQ1RIGjoWT74AWSXX725Rs3l9a2DgX2vZ1SgjdhkHrTP/nvP5b55iV06AkDzIGwnmPQoeB/j/D64WcKU4C274SJAza9oHMU52fjoOd0ld5Maz+yonlpy+d5bAI598IIKWbU8I4VCmjDixBmHqjqsxZW8Ar365nxsqtRIQJF/dpx01DsoJfYihYA0vfh6XvORcncO5eb/m6/qs8qqvg+/+DH16EIb+FATfVfuforYYv/uwck9LNuRvMOhU6nQyxblObpxK++gvMfQ7a94crX4PEjv4/e/UM58IRGbv/Qrv3d1y6k3xKN0Pplv0/Jfmw8lPYtQ3Se8GQ38Bxl0J4ZP3+bWravdNJFMUbfWLaDKVbnd/lxc4F3bdqKyrOWefvb6rqvOfe99q1bX/pJbIlDLwFTv4NtEwO7Hn58lbDorfg60egbKtTCjz7QUg6xt3uhRVTncSwLcf5Dpx2j1NNVVviBeffO3curJnhJMGuw+r1pscShDlqO3dV8umSTbz9w0ZWbC4hqWUU1w3qyHWDO5HmZyTVBuGphG3Lnbrhpe/DpgWAQOYpTskhLg0mXuv8h7r6zbo1Knq9h95v/f/g07th+wpo1cG58B1/BVz8pHORq6miDCb/wqkWGPxrOPfh2u/4cj6Cj2537i4vfcG5g/V6nTrqpZNg+RTnrjwqzrmj9JQf+rzAuetu3x9O+jUcMzSw1TwNqdoDu7Y7CaN1Zt1LDIFQUQZznoH//ce5sA8cCxkDYPa/YOsy54Zl6L1OYj6Ku/76ZAnCHJHdlR6m52xlysJ8Zv+0A49X6d4mnpuGZDG8bztiIhvwC15V7lSJbFrk1BVvWujciVVXOtvb9IbeV8Fxl0Ern2cq5z4Pn98LZ9wHp//+4O+/eye8N8ap4jlmqFN073Lez3u3lG6F6fc7VTStOsL5f3eSz7f/gm/+CknHwlUTIP24/ceUbIK3r3YuDuf/w7m7rYuda2HSaNiyBHoMd867ONep7+5+AfS6yimFhEdCedH+u/HSLVC25cBqnLh0iLSxpxpMyWaY+VdY+KZTukk61kkMx1/eaBLDXpYgTJ3trUKaNG/jvrkU2rWKYXjf9ozo244ebQPcO6Wi1Gm8LVjt1mX/5FQXFW1kX310dCto1xfa9XN/n+DUk/s/IfhgrFPtdO17TiNjTQVr4K0roTgPjrsE1s2G0k3s791yPoRHOdUDVXucqplT73F6/ey1bjZMvhnKS+DCx6HfdbBlKbx1FVSUONVF/j67NlXlTrXUorec+uleV0L3C50qGNM0bFsBhRug89kQHuznkv2zBGHqZO7aAv49/Ud+WLeTVrGRXNCrLZf0bceJmUmEhTVAdcSab2DKbc6dMDj1yb69YFK7OkmhddbhVY9U7oaXz3XuwMfO3F8vDE711MRrnOqcke9Ax0FOUtmyxGkgXDUNNi929j3mDLjgcScmf0q3Okli/WzodiGs+69Tv3zNu05D9JFSDZ3qINPoWIIwtfph3U6emP4jc9YWkBYfza+GHsvIgR2PrgrJ64XF7zgNrh0Gwal3/7zqx1dVOcwYB3OfhZSucO6j0OZ4p3qkvi6MO9fB+KGQ0F+bkKEAABvoSURBVB5+Md1pAF38Lky93enWeO2knycOXyWbnJ/2/Q8dj7caZj4Gs/7pJIVrJkFC2/o5B2MCwBKE8WthbiH/+vJHvl29g5S4aG4beizXDjrKxABO1cqn9zhdBVO7O1U4ItB/DJxy188vmFtznAbcbcvhxFuc7p1Rhz/fQ52s/grevAKOv8xJRDP/5lTdXP3G/p5E9aVgjZOMrN7fNHI25aj5mapqL09+9SPPzVxDUoso/t8FPbhucCdio44yMZSXOI21P4yH2EQY8Sz0ucbpHz7rcch+BeZPgAE3wpA7nf7g0//idEG95j3oem79nODBdD4bzrrfKa2AE9vF/wnM8waN6dkLY46QlSCamY07d3PHOwtZtLGIqwd04IGLe9Iy+ijvE7xe5+GkL/+f82TqgBvhzPsP7G5YuN6peln0jttvXZ1eQMOfqdtYOPVBFaY/4PTqOenXVrdvmj2rYjIAfLQon/s+XAYCf7usFxf1bvfzHbxe2LrU6ecPkDnEeaDK33MBqk5V0tJJsHSy0+unXT+48F9OXX1tdq6FOc86dfQnjLaLtDFBZFVMzdyuCg9/mbqc9+fncULHRP4zsp8zr7PX69T9r//W6aa54X9On3pfMYnOg2eZpzj19VEtnNLCkvec5xLCIpyqm/MegZ6X1K2Pd9IxTiIxxjRqliBC3Jw1BTz0QTbhhat5prdyfttdhH/z2v6ROStLnR1bZzoDjWWd5gxUBu5AcrOc5LHyk5+/cceT4cJ/O0+EBvPJVWNMwFiCCFHbSsv558cL6Zjzf0yN+ISoKA/8iPPTqoPzXEHfUc5DZpmnQGKHA9+kz9XODzgjWK7/1hk3p/tF/vc3xoQUSxAhxlPt5c25G5j/5dv8kVfIiNhB9fFXQo8LnQfOko45sm6kiR2h7zWH3s8YEzIsQYSQBbmFPPP+V1xT+CxjwhdSmdQNhr9OeOaQYIdmjGmCLEGEAK9XeeazBXi+e47nIqcSER2BnvEwUYNvC/xQzsaYkGUJoinzVLA753OWf/4SY3fNISayiqoelxAx7K8HH9bCGGPqyBJEU+P1Ot1Rl06ievlHtKgo5hhNYH2nK+h27i+IzPDbndkYYw6bJYimxFsNk26AlZ/giWjJtKr+fBl+KtdfM5pBndODHZ0xJsQENEGIyDDgP0A48JKqPlZje0dgApDo7nOvqk4TkUxgBeDOGM5cVb01kLE2CV/eBys/YW7mr7hx1Ykc0zaV8TcMoH1ibLAjM8aEoIAlCBEJB54FzgHygHkiMlVVc3x2uw+YpKrPi0hPYBqQ6W5bo6p9AxVfk/P9eJj7HN+0vpwbV57C8D7t+PvlvY9+gD1jjDmIQJYgBgKrVXUtgIhMBEYAvglCgb1TlLUCNgUwnqZr1efo53/kh6jB/GLLpfxhWDduO/1YxMYwMsYEUB1mcT9i7YGNPst57jpfDwLXiUgeTunhDp9tWSKyUET+KyKn+vsAERkrItkikr19+/Z6DL0R2bQI73s3soosbttzG89eeyK/GtrZkoMxJuACmSDqYhTwmqpmABcAb4hIGLAZ6Kiq/YC7gLdF5IDJkFV1vKoOUNUBqakNNFx0QyrOo+L1K9lS1YK7Iv7E67eewbDj2wQ7KmNMMxHIKqZ8wHfAngx3na+bgWEAqjpHRGKAFFXdBlS46+eLyBqgK9BsxvPW8mJ2vngJUXtKeTTxn7xy08W0aWWzkxljGk4gSxDzgC4ikiUiUcBIYGqNfXKBswBEpAcQA2wXkVS3kRsROQboAqwNYKyNSvmuYlY/fSmtStfwSruH+OevR1lyMMY0uEOWIETkYuBTVfUezhurqkdEbge+wOnC+oqqLheRcUC2qk4F7gZeFJHf4TRYj1FVFZHTgHEiUgV4gVtVdefhnVrTtDEvl92vXkZnz2o+7/IAd1wzlrAwa28wxjS8Q84oJyJvAicBk3Eu8isbIrDDFQozys38PptO066jrexg1SlP0edsGz3VGBNYtc0od8gqJlW9DugHrAFeE5E5bu+h+HqOs9mq9Hj5v0lT6T7tClLDSii+4j1LDsaYoKtTG4SqlgDvAxOBtsClwAIRuaPWA80h5Rft4cFnXuSa5b+kRVQ40bd8QfrxZwQ7LGOMqVMbxHDgRqAz8DowUFW3iUgLnIfeng5siKFr565K/vXUE/y1+t944jOI+8VUZ2IeY4xpBOrSzfVy4AlVneW7UlV3i8jNgQmreZjw8XQeq34cT9rxxI35EFomBzskY4zZpy4J4kGcB9cAEJFYIF1V16vqjEAFFupWbinh+OX/whsZTYvR71tyMMY0OnVpg3gPp6vpXtXuOnOEVJX333+bc8Ln4z3lLohLC3ZIxhhzgLokiAhVrdy74L6OClxIoe/L5Zu4fNuzlMa0o8Wp1s5vjGmc6pIgtrsN1QCIyAhgR+BCCm0VnmoWTX2WHmG5tLjgEYi0J6SNMY1TXdogbgXeEpFnAMEZofWGgEYVwt6YuZybKt6kOK0/rXpdFuxwjDHmoA6ZIFR1DTBYROLc5bKARxWitpWW4539L1KlGC75J9iQ3caYRqxOo7mKyIXAcUDM3nkIVHVcAOMKSS9NncndfEpZt8uJa98/2OEYY0ytDtkGISL/B1yNM5mPAFcCnQIcV8hZmldMr5VPEBYeRtwFDwc7HGOMOaS6NFKfrKo3AIWq+hDOwH1dAxtWaFFVJk5+j4vD5+I96Q5oVXNiPWOMaXzqkiDK3d+7RaQdUIUzHpOpo48X53NlwXPsjk4j+vS7gh2OMcbUSV3aID4WkUTgn8ACnHkbXgxoVCFkd6WH5Z88w/CwNXiHPQdRLYMdkjHG1EmtCcKdH3qGqhYBk0XkEyBGVYsbJLoQ8OYX33F71WuUtD2JhD6jgh2OMcbUWa1VTO4scs/6LFdYcqi7jQW76D7vPqLClISr/w/CAjnDqzHG1K+6XLFmiMjlItZp/3B9/e6TnBa2mD2n3w+tM4MdjjHGHJa6JIhf4gzOVyEiJSJSKiIlAY6rycteuoxLtz5DfsIJJJ72q2CHY4wxh60uT1Lb1KKHyeOpxvPRnUSJh5Rrx1vVkjGmSarLjHKn+VtfcwIhs9/cj57nFM88VvS+lx7pXYIdjjHGHJG6dHP9vc/rGGAgMB8481AHisgw4D9AOPCSqj5WY3tHYAKQ6O5zr6pOc7f9CbgZZ/6J36jqF3WINeiKt22k19K/8mNkD7qP+P2hDzDGmEaqLlVMF/sui0gH4MlDHSci4Tg9oM4B8oB5IjJVVXN8drsPmKSqz4tIT2AakOm+Hokz/lM74CsR6aqq1XU8r+BQJf+tX3GsVhJ+6XNIeJ2GujLGmEbpSCrH84AeddhvILBaVde6kwxNBEbU2EeBBPd1K2CT+3oEMNHtVrsOWO2+X6OWP3cyPYtnMbPdLRzb84Rgh2OMMUelLm0QT+NcyMFJKH1xnqg+lPY4c0fslQcMqrHPg8CXInIH0BI42+fYuTWOPWAAIxEZC4wF6NixYx1CCqzy/z3HRk1j4DX3BzsUY4w5anUpQWTjtDnMB+YAf1TV6+rp80cBr6lqBnAB8Ib79HadqOp4VR2gqgNSU1PrKaQj492xlmPL5rMg+SJax7cIaizGGFMf6lJJ/j5Qvrf+X0TCRaSFqu4+xHH5QAef5Qx3na+bgWEAqjpHRGKAlDoe26hsmfUS6SrEDrTJ9owxoaFOT1IDsT7LscBXdThuHtBFRLJEJAqn0XlqjX1ygbMARKQHTi+p7e5+I0UkWkSygC7AD3X4zOCo9hC34l1maV9O7tcr2NEYY0y9qEuCiPGdZtR9fcg6FFX1ALcDXwArcHorLReRcSIy3N3tbuAWEVkMvAOMUcdyYBKQA3wO/Lox92Dy/vQVCVU7yGkzgrho67lkjAkNdbma7RKRE1R1AYCI9Af21OXN3WcaptVY94DP6xxgyEGOfRR4tC6fE2wl371MlSbQ9sRLgh2KMcbUm7okiDuB90RkE86Uo21wpiA1AKVbid84g5e953P1cRnBjsYYY+pNXR6Umyci3YFu7qpVqloV2LCaDl38DuFazZqMy2jVIjLY4RhjTL05ZBuEiPwaaKmqy1R1GRAnIjY8KYAqlfMm8IO3G/36nRjsaIwxpl7VpZH6FndGOQBUtRC4JXAhNSG5c4guXsuk6jM4p2d6sKMxxph6VZcEEe47WZA7xlJU4EJqOnTB6+wiloKO55McFx3scIwxpl7VpZH6c+BdEXnBXf4l8FngQmoiyovR5R8yxTOEM3tnBjsaY4ypd3VJEH/EGe/oVnd5CU5PpuZt2WTCPOVM8g7lxePsz2GMCT2HrGJSVS/wPbAeZ0TVM3EefGveFrzO2rBMojr0Jy0hJtjRGGNMvTtoCUJEuuIMpjcK2AG8C6CqZzRMaI3YlqWwaSGvV93AsF7tgh2NMcYERG1VTCuB2cBFqroaQER+1yBRNXbzX8MjUXxYfQrTjrfqJWNMaKqtiukyYDPwjYi8KCJn4TxJ3bwVboAFrzMj6gwyO2TQPjH20McYY0wTdNAEoapTVHUk0B34BmfIjTQReV5Ezm2oABudb/6KlzD+UnwR51vpwRgTwurSSL1LVd9256bOABbi9GxqfrYsgyXvsrT91Wwh2RKEMSakHdac1Kpa6M7idlagAmrUZoxDYxJ4uOg8jm+fQKfklsGOyBhjAuawEkSztuE7+OkL8nreSvY2uHZQp2BHZIwxAWUJoi5UYfpfIL4t/y4ZSnxMBCP6WvdWY0xoswRRF6umQd4PlAy6m49ziriyfwdaRNnMccaY0GYJ4lCqPTBjHCR34bU9p+DxKtefZNVLxpjQZwniUBa/A9tX4jnjPt6al8+pXVLISrHGaWNM6LMEUZuqPTDzb9C+P9O9A9laUsENJ2UGOypjjGkQAU0QIjJMRFaJyGoRudfP9idEZJH786OIFPlsq/bZNjWQcR7UDy9CST6c/RCvz82lfWIsZ3ZPC0ooxhjT0ALW0upOLPQscA6QB8wTkamqmrN3H1X9nc/+dwD9fN5ij6r2DVR8dTLvJcg6nZ9a9GXO2ln8YVg3wsNstBFjTPMQyBLEQGC1qq5V1UpgIjCilv1HAe8EMJ7D46mE4o3QcTBvzN1AVHgYVw/oEOyojDGmwQQyQbQHNvos57nrDiAinYAs4Guf1TEiki0ic0XkkoMcN9bdJ3v79u31FbejJA/US3lcBh8syOei3m1tWlFjTLPSWBqpRwLvq2q1z7pOqjoAuAZ4UkSOrXmQO+zHAFUdkJqaWr8RFeUCMGtbLGUVHuvaaoxpdgKZIPIB3zqZDHedPyOpUb2kqvnu77XATH7ePhF4hRsAmLACerVvRd8OiQ368cYYE2yBTBDzgC4ikiUiUThJ4IDeSCLSHWgNzPFZ11pEot3XKcAQIKfmsQFVlItKOHN3RHP9SZ0QscZpY0zzErBeTKrqEZHbgS+AcOAVVV0uIuOAbFXdmyxGAhNVVX0O7wG8ICJenCT2mG/vpwZRtIGC8DTiW8QwvI+Nu2SMaX4COqCQqk4DptVY90CN5Qf9HPcd0CuQsR1SUS7rPMmce3w6MZHhQQ3FGGOCobE0Ujc6WriBtZ4UMlq3CHYoxhgTFJYg/Knag5RtIU9TaJMQE+xojDEmKCxB+FOcB8BGTSO9lSUIY0zzZAnCH7eLq5UgjDHNmSUIf4qcBLFR0yxBGGOaLUsQ/hRtwCORlEQmkRBrM8cZY5onu/r5U5TLzoh00lu0sAfkjDHNlpUg/CncwGZJJd2ql4wxzZglCH+KclnvTaWN9WAyxjRjVsVUU0UZ7N7B6uoka6A2xjRrVoKoqdiZwmJ9dYpVMRljmjVLEDXtewbCqpiMMc2bJYia3ImCNmqalSCMMc2aJYiaijbgCYthBwlWgjDGNGuWIGoq2kBxdFtEhLR4m4PaGNN8WYKoqXAD2yPSSW4ZTWS4/XmMMc2XXQFrKtrgNlBb6cEY07xZgvC1pwjKi1nnsVFcjTHGEoQvtwfTqorW1oPJGNPsWYLw5SaIleX2FLUxxgQ0QYjIMBFZJSKrReReP9ufEJFF7s+PIlLks220iPzk/owOZJz7FNlDcsYYs1fAxmISkXDgWeAcIA+YJyJTVTVn7z6q+juf/e8A+rmvk4C/AAMABea7xxYGKl4AinKpjoyjqDzOEoQxptkLZAliILBaVdeqaiUwERhRy/6jgHfc1+cB01V1p5sUpgPDAhiro3ADZbHtALEqJmNMsxfIBNEe2OiznOeuO4CIdAKygK8P51gRGSsi2SKSvX379qOPuCiXwqi2AKRbCcIY08w1lkbqkcD7qlp9OAep6nhVHaCqA1JTU48uAlUo2sDWsHRaRIUTH20joRtjmrdAJoh8oIPPcoa7zp+R7K9eOtxj68eeQqgsY6M3lTYJMTbVqDGm2QtkgpgHdBGRLBGJwkkCU2vuJCLdgdbAHJ/VXwDnikhrEWkNnOuuC5zC9QCsqUqyZyCMMYYA9mJSVY+I3I5zYQ8HXlHV5SIyDshW1b3JYiQwUVXV59idIvIwTpIBGKeqOwMVK7DvGYicPa1p084ShDHGBLSiXVWnAdNqrHugxvKDBzn2FeCVgAVXk/sMxJJdrbjaShDGGNNoGqmDrygXb0wihdWxtEmwgfqMMcYSxF6FG6homQFgD8kZYwyWIPYr2kBpbDsAa6Q2xhgsQThUoSiXHZFtACtBGGMMWIJwlG0DTzlbSCNMIDXO2iCMMcYSBOzr4rrem0pKXDQRNtWoMcZYggD2dXH9saK1VS8ZY4zLEgTsSxA5uxOtgdoYY1yWIAAKN0CLFNaXYMN8G2OMyxIEOA/JJXakpNxjVUzGGOOyBAFQtIHdLZzpJqwEYYwxDksQXi8UbaQ42nlIzkoQxhjjsARRuhm8VWwPTwfsKWpjjNnLEkR8W7hrJQsSzgKsBGGMMXtZgggLg4S25O6OJC46gjibatQYYwBLEPtsLSkn3Yb5NsaYfSxBuLaUlFv1kjHG+LAE4dpaXG4N1MYY48MSBOD1KttKK+wZCGOM8WEJAtixqwKPV62KyRhjfFiCALYWVwD2DIQxxvgKaIIQkWEiskpEVovIvQfZ5yoRyRGR5SLyts/6ahFZ5P5MDWScW0rKARtmwxhjfAWs07+IhAPPAucAecA8EZmqqjk++3QB/gQMUdVCEUnzeYs9qto3UPH52pcgrIrJGGP2CWQJYiCwWlXXqmolMBEYUWOfW4BnVbUQQFW3BTCeg9paXE54mJBiU40aY8w+gUwQ7YGNPst57jpfXYGuIvI/EZkrIsN8tsWISLa7/hJ/HyAiY919srdv337EgW4pKSc1LprwMDni9zDGmFAT7HElIoAuwFAgA5glIr1UtQjopKr5InIM8LWILFXVNb4Hq+p4YDzAgAED9EiD2FJcTrpVLxljzM8EsgSRD3TwWc5w1/nKA6aqapWqrgN+xEkYqGq++3stMBPoF6hAt5SU08aG2TDGmJ8JZIKYB3QRkSwRiQJGAjV7I03BKT0gIik4VU5rRaS1iET7rB8C5BAgW4vLrQeTMcbUELAqJlX1iMjtwBdAOPCKqi4XkXFAtqpOdbedKyI5QDXwe1UtEJGTgRdExIuTxB7z7f1Un3ZVeCit8FgVkzHG1BDQNghVnQZMq7HuAZ/XCtzl/vju8x3QK5Cx7VXh8XJxn3Yc365VQ3ycMcY0GcFupA66pJZRPD0qYM0bxhjTZNlQG8YYY/yyBGGMMcYvSxDGGGP8sgRhjDHGL0sQxhhj/LIEYYwxxi9LEMYYY/yyBGGMMcYvcR5mbvpEZDuw4SjeIgXYUU/hNCV23s2LnXfzUpfz7qSqqf42hEyCOFoikq2qA4IdR0Oz825e7Lybl6M9b6tiMsYY45clCGOMMX5ZgthvfLADCBI77+bFzrt5OarztjYIY4wxflkJwhhjjF+WIIwxxvjV7BOEiAwTkVUislpE7g12PIEkIq+IyDYRWeazLklEpovIT+7v1sGMsb6JSAcR+UZEckRkuYj81l0f6ucdIyI/iMhi97wfctdnicj37vf9XXe++JAjIuEislBEPnGXm8t5rxeRpSKySESy3XVH/F1v1glCRMKBZ4HzgZ7AKBHpGdyoAuo1YFiNdfcCM1S1CzDDXQ4lHuBuVe0JDAZ+7f4bh/p5VwBnqmofoC8wTEQGA38HnlDVzkAhcHMQYwyk3wIrfJaby3kDnKGqfX2efzji73qzThDAQGC1qq5V1UpgIjAiyDEFjKrOAnbWWD0CmOC+ngBc0qBBBZiqblbVBe7rUpyLRntC/7xVVcvcxUj3R4Ezgffd9SF33gAikgFcCLzkLgvN4LxrccTf9eaeINoDG32W89x1zUm6qm52X28B0oMZTCCJSCbQD/ieZnDebjXLImAbMB1YAxSpqsfdJVS/708CfwC87nIyzeO8wbkJ+FJE5ovIWHfdEX/XI+o7OtN0qaqKSEj2exaROGAycKeqljg3lY5QPW9VrQb6ikgi8CHQPcghBZyIXARsU9X5IjI02PEEwSmqmi8iacB0EVnpu/Fwv+vNvQSRD3TwWc5w1zUnW0WkLYD7e1uQ46l3IhKJkxzeUtUP3NUhf957qWoR8A1wEpAoIntvDEPx+z4EGC4i63GqjM8E/kPonzcAqprv/t6Gc1MwkKP4rjf3BDEP6OL2cIgCRgJTgxxTQ5sKjHZfjwY+CmIs9c6tf34ZWKGq//bZFOrnneqWHBCRWOAcnPaXb4Ar3N1C7rxV9U+qmqGqmTj/n79W1WsJ8fMGEJGWIhK/9zVwLrCMo/iuN/snqUXkApw6y3DgFVV9NMghBYyIvAMMxRkCeCvwF2AKMAnoiDNc+lWqWrMhu8kSkVOA2cBS9tdJ/xmnHSKUz7s3ToNkOM6N4CRVHScix+DcWScBC4HrVLUieJEGjlvFdI+qXtQczts9xw/dxQjgbVV9VESSOcLverNPEMYYY/xr7lVMxhhjDsIShDHGGL8sQRhjjPHLEoQxxhi/LEEYY4zxyxKEMYdBRKrdkTL3/tTbIH8ikuk70q4xwWZDbRhzePaoat9gB2FMQ7AShDH1wB2H/x/uWPw/iEhnd32miHwtIktEZIaIdHTXp4vIh+58DYtF5GT3rcJF5EV3Docv3aegjQkKSxDGHJ7YGlVMV/tsK1bVXsAzOE/nAzwNTFDV3sBbwFPu+qeA/7rzNZwALHfXdwGeVdXjgCLg8gCfjzEHZU9SG3MYRKRMVeP8rF+PM0HPWndwwC2qmiwiO4C2qlrlrt+sqikish3I8B3uwR2OfLo7sQsi8kcgUlUfCfyZGXMgK0EYU3/0IK8Ph+/4QNVYO6EJIksQxtSfq31+z3Fff4czqijAtTgDB4Iz9eNtsG9in1YNFaQxdWV3J8Ycnlh3lra9PlfVvV1dW4vIEpxSwCh33R3AqyLye2A7cKO7/rfAeBG5GaekcBuwGWMaEWuDMKYeuG0QA1R1R7BjMaa+WBWTMcYYv6wEYYwxxi8rQRhjjPHLEoQxxhi/LEEYY4zxyxKEMcYYvyxBGGOM8ev/Ax/uoeHhjYonAAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = Transformer(50).cuda()\n",
        "train_network(model, train_data, valid_data, batch_size = 100, num_epochs=50, learning_rate=2e-4)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "pUvLteKUI7jI",
        "outputId": "54c5048e-c088-4394-f033-043b0ebedcdf"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1; Loss 0.603379; Train Acc 0.829408; Val Acc 0.817939\n",
            "Epoch 2; Loss 0.230217; Train Acc 0.862420; Val Acc 0.856914\n",
            "Epoch 3; Loss 0.324370; Train Acc 0.877944; Val Acc 0.868660\n",
            "Epoch 4; Loss 0.189075; Train Acc 0.888473; Val Acc 0.872931\n",
            "Epoch 5; Loss 0.267516; Train Acc 0.901856; Val Acc 0.888414\n",
            "Epoch 6; Loss 0.529546; Train Acc 0.891863; Val Acc 0.872397\n",
            "Epoch 7; Loss 0.022195; Train Acc 0.914347; Val Acc 0.900160\n",
            "Epoch 8; Loss 0.019145; Train Acc 0.925232; Val Acc 0.905499\n",
            "Epoch 9; Loss 0.089620; Train Acc 0.934868; Val Acc 0.899092\n",
            "Epoch 10; Loss 0.021159; Train Acc 0.923804; Val Acc 0.899092\n",
            "Epoch 11; Loss 0.028659; Train Acc 0.952534; Val Acc 0.906567\n",
            "Epoch 12; Loss 0.023044; Train Acc 0.947894; Val Acc 0.907101\n",
            "Epoch 13; Loss 0.021686; Train Acc 0.912919; Val Acc 0.873999\n",
            "Epoch 14; Loss 0.033543; Train Acc 0.938615; Val Acc 0.894821\n",
            "Epoch 15; Loss 0.013091; Train Acc 0.943790; Val Acc 0.895889\n",
            "Epoch 16; Loss 0.174627; Train Acc 0.972698; Val Acc 0.902296\n",
            "Epoch 17; Loss 1.301438; Train Acc 0.890079; Val Acc 0.853177\n",
            "Epoch 18; Loss 0.007423; Train Acc 0.965382; Val Acc 0.911906\n",
            "Epoch 19; Loss 0.003446; Train Acc 0.971271; Val Acc 0.904431\n",
            "Epoch 20; Loss 0.011214; Train Acc 0.966274; Val Acc 0.893753\n",
            "Epoch 21; Loss 0.007819; Train Acc 0.981442; Val Acc 0.901228\n",
            "Epoch 22; Loss 0.020417; Train Acc 0.983226; Val Acc 0.903897\n",
            "Epoch 23; Loss 0.014880; Train Acc 0.986974; Val Acc 0.907101\n",
            "Epoch 24; Loss 0.002032; Train Acc 0.989115; Val Acc 0.906033\n",
            "Epoch 25; Loss 0.006369; Train Acc 0.992505; Val Acc 0.906567\n",
            "Epoch 26; Loss 0.018721; Train Acc 0.991613; Val Acc 0.904431\n",
            "Epoch 27; Loss 0.008066; Train Acc 0.992327; Val Acc 0.900160\n",
            "Epoch 28; Loss 0.003842; Train Acc 0.989293; Val Acc 0.903364\n",
            "Epoch 29; Loss 0.006900; Train Acc 0.990542; Val Acc 0.906033\n",
            "Epoch 30; Loss 0.002441; Train Acc 0.991435; Val Acc 0.902296\n",
            "Epoch 31; Loss 0.000854; Train Acc 0.990364; Val Acc 0.903897\n",
            "Epoch 32; Loss 0.005926; Train Acc 0.995717; Val Acc 0.911372\n",
            "Epoch 33; Loss 0.050634; Train Acc 0.981799; Val Acc 0.902830\n",
            "Epoch 34; Loss 0.608714; Train Acc 0.978408; Val Acc 0.905499\n",
            "Epoch 35; Loss 0.002916; Train Acc 0.973947; Val Acc 0.899092\n",
            "Epoch 36; Loss 0.003437; Train Acc 0.989472; Val Acc 0.901762\n",
            "Epoch 37; Loss 0.000898; Train Acc 0.993933; Val Acc 0.902296\n",
            "Epoch 38; Loss 0.000550; Train Acc 0.986974; Val Acc 0.896423\n",
            "Epoch 39; Loss 0.001982; Train Acc 0.988401; Val Acc 0.899092\n",
            "Epoch 40; Loss 0.000941; Train Acc 0.987687; Val Acc 0.894287\n",
            "Epoch 41; Loss 0.000804; Train Acc 0.994111; Val Acc 0.902296\n",
            "Epoch 42; Loss 0.000451; Train Acc 0.984654; Val Acc 0.896957\n",
            "Epoch 43; Loss 0.000882; Train Acc 0.996610; Val Acc 0.902830\n",
            "Epoch 44; Loss 0.001021; Train Acc 0.995896; Val Acc 0.901762\n",
            "Epoch 45; Loss 0.000618; Train Acc 0.992148; Val Acc 0.897491\n",
            "Epoch 46; Loss 0.002789; Train Acc 0.995539; Val Acc 0.902830\n",
            "Epoch 47; Loss 0.001559; Train Acc 0.994111; Val Acc 0.901762\n",
            "Epoch 48; Loss 0.001547; Train Acc 0.985546; Val Acc 0.893219\n",
            "Epoch 49; Loss 0.001783; Train Acc 0.995360; Val Acc 0.904965\n",
            "Epoch 50; Loss 0.002044; Train Acc 0.972877; Val Acc 0.884143\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3de5zcdX3v8ddn7nvf7CUXciEXghAURSOI4AEFNdgK3qpQtNpqUVs9erS22ItYak/v1nrE06Ki1runaqU1lnurlSAEEIRcIAkJSUiym83ed2bn9jl//H6/2cnu7O5skt/8JvP7PB+PfezMb34z851kdt7zvYuqYowxJrwiQRfAGGNMsCwIjDEm5CwIjDEm5CwIjDEm5CwIjDEm5CwIjDEm5CwITGiIyI9F5J2n+lxjTndi8whMPRORsbKrzcAkUHCvv1dVv1H7Up0cEWkHbgbeBHQBR4B/Az6lqkeDLJsJJ6sRmLqmqq3eD/As8PqyY6UQEJFYcKWsnogkgHuA84BNQDtwMTAAXHgCj3davG5T3ywIzGlJRC4XkQMi8gcichj4sogsEpF/F5F+ERl0L68ou89/ish73MvvEpH/FpG/dc99RkSuOsFz14jIT0RkVETuFpFbROTrsxT9N4BVwBtVdZuqFlW1T1X/TFU3u4+nInJW2eN/RUQ+Ncfr3i4iv1p2fsz9N3ixe/1lInK/iAyJyGMicvnJ/vubxmJBYE5nS3GaVs4EbsB5P3/Zvb4KSAOfm+P+FwE7gR7gr4EviYicwLnfBB4EuoFPAu+Y4zmvBP5DVcfmOGc+01/3t4Drym5/LXBUVR8RkeXAj4BPuff5PeB7ItJ7Es9vGowFgTmdFYGbVHVSVdOqOqCq31PVCVUdBf4cuGyO++9T1S+oagH4KrAMWLKQc0VkFfBS4BOqmlXV/wZun+M5u4FDC3uZMxz3unGC6GoRaXZv/3WccAB4O7BZVTe7tY+7gK3A606yDKaBWBCY01m/qma8KyLSLCL/JCL7RGQE+AnQKSLRWe5/2LugqhPuxdYFnnsGcKzsGMD+Oco8gBMiJ+O4162qu4DtwOvdMLgaJxzAqTX8mtssNCQiQ8Clp6AMpoFYR5M5nU0f8vZR4HnARap6WEReBDwKzNbccyocArpEpLksDFbOcf7dwKdEpEVVx2c5ZwJnhJRnKXCg7HqloX5e81AE2OaGAzih9DVV/e15XocJMasRmEbShtMvMCQiXcBNfj+hqu7DaWr5pIgkRORi4PVz3OVrOB/O3xORc0QkIiLdIvKHIuI11/wC+HURiYrIJuZu3vJ8G3gN8H6magMAX8epKbzWfbyU2+G8ouKjmFCyIDCN5DNAE3AUeAD4jxo97/VMDQH9FPAdnPkOM6jqJE6H8Q7gLmAEp6O5B/i5e9qHcMJkyH3sf52vAKp6CNgCvNx9fu/4fuAa4A+BfpwQ+hj2t2/K2IQyY04xEfkOsENVfa+RGHMq2LcCY06SiLxURNa5zTybcL6Bz/st3ph6YZ3Fxpy8pcD3cYaGHgDer6qPBlskY6pnTUPGGBNy1jRkjDEhd9o1DfX09Ojq1auDLoYxxpxWHn744aOqWnFpkdMuCFavXs3WrVuDLoYxxpxWRGTfbLdZ05AxxoScBYExxoScBYExxoScBYExxoScBYExxoScBYExxoScBYExxoScBYEJxH07+zgwODH/icYY31kQmEB84BuP8NX79wZdDGMMFgQmAKrKeLbARLYQdFGMMVgQmABM5ovH/TbGBMuCwNRcJufUBCwIjKkPFgSm5tJeEOSsaciYemBBYGouk7OmIWPqiQWBqbmppiGrERhTDywITM2lrY/AmLpiQWBqzqsReE1ExphgWRCYmpss9RFY05Ax9cCCwNRcqY/AagTG1AULAlNz1kdgTH3xLQhE5DYR6RORJ2a5/XoReVxEfiki94vIC/0qi6kvGWsaMqau+Fkj+AqwaY7bnwEuU9UXAH8G3OpjWUwdsZnFxtQX34JAVX8CHJvj9vtVddC9+gCwwq+ymPriNQ1l80VUNeDSGGPqpY/g3cCPgy6EqY3ypSWsVmBM8GJBF0BEXokTBJfOcc4NwA0Aq1atqlHJjF8yZR/+k/kiqXg0wNIYYwKtEYjI+cAXgWtUdWC281T1VlXdqKobe3t7a1dA44vMcTUC6zA2JmiBBYGIrAK+D7xDVZ8Kqhym9tJlG9LYXAJjgudb05CIfAu4HOgRkQPATUAcQFX/EfgE0A18XkQA8qq60a/ymPpxfNOQ1QiMCZpvQaCq181z+3uA9/j1/KZ+lTcN2XpDxgSvXkYNmRDJ2KghY+qKBYGpuUyuQEScy9Y0ZEzwLAhMzWVyRTqa4oDVCIypBxYEpuYyucJUEFgfgTGBsyAwNZcuDwJrGjImcBYEpuYyuSIdzQnAmoaMqQcWBKbmJo+rEVgQGBM0CwJTc5l8gY4mZwpL+QJ0xphgWBCYmsoXiuQKajUCY+qIBYGpKW95ifaUBYEx9cKCwNSUN6u4KRElGYtY05AxdcCCwNSUt/JoKhYlFY9ajcCYOmBBYGrKmzeQ8moENo/AmMBZEJia8lYbTcUiJOMRm1lsTB2wIDA15fURpOJRkjFrGjKmHlgQmJpKT+8stqYhYwJnQWBqaqppyAsCqxEYEzQLAlNTU01DEadpyPoIjAmcBYGpqXR5H0HcmoaMqQcWBKamJo/rLLamIWPqgQWBqalSH4HXNGRBYEzgLAhMTWWm1QgytsSEMYHzLQhE5DYR6RORJ2a5XUTksyKyS0QeF5EX+1UWUz/SuQKxiBCPRmyJCWPqhJ81gq8Am+a4/SpgvftzA/B/fSyLqROZXJFUPApgi84ZUyd8CwJV/QlwbI5TrgH+WR0PAJ0issyv8pj6kMkXSMWdt50zashqBMYELcg+guXA/rLrB9xjM4jIDSKyVUS29vf316Rwxh+ZbIFkzKsRRMkXlXzBwsCYIJ0WncWqequqblTVjb29vUEXx5yETL5AU2KqaQgga0FgTKCCDIKDwMqy6yvcY6aBOX0EbtOQGwQ2u9iYYAUZBLcDv+GOHnoZMKyqhwIsj6mBTK5AymsacjuNrZ/AmGDF/HpgEfkWcDnQIyIHgJuAOICq/iOwGXgdsAuYAH7Tr7KY+pHOFWhNOm+7Uo3AlpkwJlC+BYGqXjfP7Qr8rl/Pb+pTJleku2WqsxisRmBM0E6LzmLTOCZzBesjMKbOWBCYmkrnClMTytxAyFjTkDGBsiAwNZXJFWhyg8ALBKsRGBMsCwJTUxWHj1qNwJhAWRCYmlFVd4kJ6yw2pp5YEJiamcwXUeW4Reec41YjMCZIFgSmZiZLm9Ic31lsfQTGBMuCwNSMNzpoqo/AmoaMqQcWBKZm0lk3CGLWNGRMPbEgMDXj1Qimrz5qTUPGBMuCwNRM+cb1ALFohGhErGnImIBZEJiaKW1c7zYNgbtdpTUNGRMoCwJTM2kvCBLTg8BqBMYEyYLA1MxkhRpBKh4t1RSMMcGwIDA1M72PAKxGYEw9sCAwNVNqGoqXNw1FbdSQMQGzIDA14zUBNZUHQdw6i40JmgWBqZnMtCUmwJqGjKkHFgSmZrwagTeRzLkctSAwJmAWBKZmMrkCyViESERKx2wegTHBsyAwNZMp26bSk4xHrLPYmID5GgQisklEdorILhG5scLtq0TkPhF5VEQeF5HX+VkeE6zy3ck81jRkTPB8CwIRiQK3AFcBG4DrRGTDtNP+GPiuql4AXAt83q/ymOClK9UIrGnImMD5WSO4ENilqntUNQt8G7hm2jkKtLuXO4DnfCyPCVj5xvUeGzVkTPD8DILlwP6y6wfcY+U+CbxdRA4Am4EPVnogEblBRLaKyNb+/n4/ympqIJMvkpwWBLbEhDHBC7qz+DrgK6q6Angd8DURmVEmVb1VVTeq6sbe3t6aF9KcGplcgVRseh9BxN3LWAMqlTHGzyA4CKwsu77CPVbu3cB3AVR1C5ACenwskwlQJlcobUrjScajqEKuYEFgTFD8DIKHgPUiskZEEjidwbdPO+dZ4AoAETkXJwis7adBOTWCmX0EYNtVGhMk34JAVfPAB4A7gO04o4OeFJGbReRq97SPAr8tIo8B3wLepdZG0LAqDx/1gsA6jI0JSszPB1fVzTidwOXHPlF2eRtwiZ9lMPWj8vBR57oFgTHBCbqz2ITIbDOLYWrTGmNM7VkQmJqZzBUrTigDqxEYEyQLAlMThaKSLVReYgIsCIwJkgWBqYlKm9JAWY3AmoaMCYwFgamJTIVtKqGsj8BqBMYExoLA1EQmP3PjephqGrJlJowJjgWBqYl0tnKNIGU1AmMCZ0FgamLWpiHrLDYmcBYEpia8JSRmHz5qTUPGBKWqIBCRFm9VUBE5W0SuFpG4v0UzjSTjbkc5c/VRt0Zg21UaE5hqawQ/AVIishy4E3gH8BW/CmUaj9dHMHP1UesjMCZo1QaBqOoE8Cbg86r6a8B5/hXLNJrMLE1Diag1DRkTtKqDQEQuBq4HfuQei85xvjHHmWoaOv5tE4kIiahtV2lMkKoNgg8DHwd+4C4lvRa4z79imUaTLo0amvmWS8Yi1kdgTICqWoZaVf8L+C8At9P4qKr+Tz8LZhqLt4REKjGzIpmMR6xpyJgAVTtq6Jsi0i4iLcATwDYR+Zi/RTONpDSPIFYhCGJRaxoyJkDVNg1tUNUR4A3Aj4E1OCOHjKlKJlckIhCPyozbnBqBBYExQak2COLuvIE3ALerag6wLSVN1dK5Ak3xKCIVgiAWtbWGjAlQtUHwT8BeoAX4iYicCYz4VSjTeCrtTuZJxqxGYEyQqu0s/izw2bJD+0Tklf4UyTSiTIXdyTzOqCGrERgTlGo7iztE5NMistX9+Tuc2oExVcnkCqVZxNMl49ZZbEyQqm0aug0YBd7q/owAX/arUKbxZNw+gkqsaciYYFUbBOtU9SZV3eP+/Cmwdr47icgmEdkpIrtE5MZZznmriGwTkSdF5JsLKbw5fWTy8/URWNNQvVFVnh2YCLoYpgaqDYK0iFzqXRGRS4D0XHcQkShwC3AVsAG4TkQ2TDtnPc6M5UtU9TycGcy+2N0/xj/9125GMjm/nsLMwekjmKVpKBa1mcV16O7tfVz+t/dxaHjOP3XTAKoNgvcBt4jIXhHZC3wOeO8897kQ2OXWILLAt4Frpp3z28AtqjoIoKp9VZd8gZ4+Mspf/HiHfcMJSDo7R9OQzSOoS/uPTVBUODycCbooxmdVBYGqPqaqLwTOB85X1QuAV81zt+XA/rLrB9xj5c4GzhaRn4nIAyKyqdIDicgNXkd1f39/NUWeYVlHE2Bv6qBk8gWS1jR0WhlKO7Xn4bTVohvdgnYoU9URd4YxwEdOwfPHgPXA5cB1wBdEpLPC896qqhtVdWNvb+8JPdGyjhQAh0YsCIIwmStWXF4CbImJejU0kQUsCMLgZLaqnDlF9HgHgZVl11e4x8odwJ2prKrPAE/hBMMp192aJBYRDlt7ZyDSucKsfQSpeIRsvoiqTVavJ4MTViMIi5MJgvn+ah8C1ovIGhFJANcCt087519xagOISA9OU9GekyjTrKIRYUl7ikPWNBSIuYeP2gb29ahUI5iwIGh0c84sFpFRKn/gC9A0131VNS8iHwDuwNnE5jZ3L4Obga2qert722tEZBtQAD6mqgMn8DqqsrQjZX0EAVDVeZeYALf5aJZzTO0NWY0gNOYMAlVtO5kHV9XNwOZpxz5Rdllx+hpORX/DvJZ2pNj+nC2RVGu5glLUypvSQPm+xQUgXsOSmbkMujWCIQuChncyTUOnnWVu05C1RdfW1O5k1jR0OrEaQXiEKgiWdqRI5wqMpPNBFyVUJucNAtvAvt5k80XGJp2/EwuCxheqIPDmEhwasZFDtVTauH6eIMjY7OK6Uf7hb53FjS9UQbDUm0swVPsO4+eGwhs+XtPQ7DOLrWmo3ngjhpoTUasRhECogqA0qazGI4ce2nuMl//lvew8PFrT560Xpf2KZ11ryJqG6o03h+DM7hYLghAIVRD0tiWJCDWfVPb4gWEAnj0WznWOMlX3EViNoF54I4bO7GomnStYSDe4UAVBPBqhty1Z8xrB7v4xAI6NT9b0eetFJu/1Ecy++ihgK5DWEa9f4MyeZue61QoaWqiCAGBpRxOHa7ze0O4+LwjC+ceUzs5dI0jFrWmo3ng1gtXdzkaEIxYEDS10QbAsgGUmdvePA+GtEXgf8LM2DcWtRlBvBidyxKPCGZ3OSDurETS20AVBrZeZGJ7IcXTMCYCB8WzNnreeVN9HYDWCejE0kaWzOUFHU9y9bkHQyEIXBMs6UoxN5hmt0U5lu4+OlS4PhjQIvKahufYsBussridDEzk6m+J0ukFgNYLGFrog8OYS1KpW4PUPrOpq5lhIg6DqzmILgroxOJFlUVmNwIKgsYUuCEqzi2sVBP3jxKPCC1d2WtPQLBvTxKOCyNRSFCZ4QxM5OpvjtFsQhEIIg6DGNYL+MVZ3t7C4LRnapqFMrkgiGiESqbyXkYi421VajaBeeDWCaERoS8asj6DBhS4IFrcngdrVCPb0j7Gut5WulgTj2ULp23GYZObYncxj21XWD1VlKO3UCAA6muM2fLTBhS4IkrEoPa0JDtdg4blcoci+gQnWLW6hqyUBTI3PDpO5NqXx2Ab29SOdK5DNF+lsdt6zHU1xaxpqcKELAnD6CWpRI3j22AT5orKut5VF7h/VwJgFQSXJeMTmEdQJb52hRV6NwIKg4YUyCGo1l8AbMbSut5XuVicIwjhyKD3HfsUeaxqqH15fVnmNwHYpa2yhDIJlHbWZXezNKF7bG/amoWIVfQTWNFQvvG//Xh9BZ7PVCBrdnHsWN6qlHSmG0zkmsnmaE/79E+zuH2NJe5K2VJx8wdkeM6xNQ8l5agSpeNQ2pqkT3pcVrzmz3W0aUlVEKo/8Mqe30NYIwP8hpLv7x1jb0wo41etoRELZNJTJF62z+DRSqY8gmy9aUDcwX4NARDaJyE4R2SUiN85x3ptFREVko5/l8SxtdyaV+RkEqsruvjHWLXZWb4xEhEXNcY6FsWkoW6CpqqYh+6CpB0Pul5UOr2moyakZWPNQ4/ItCEQkCtwCXAVsAK4TkQ0VzmsDPgT83K+yTFeLncqOjmUZyeRZ19taOraoOcGxMDYN5asZPhq1UUN1YiidozkRLS39YctMND4/awQXArtUdY+qZoFvA9dUOO/PgL8CarYkaGm9IR/3JfA2oykPgq6WRDhrBLnCrMtLeJJxaxqqF96sYo8FQePzMwiWA/vLrh9wj5WIyIuBlar6o7keSERuEJGtIrK1v7//pAuWikdZ1BznkI9bVpaCYPFUEHS3JkLZR5DOFmhKVNNHYDWCeuCtM+SZWoo6fO/dsAiss1hEIsCngY/Od66q3qqqG1V1Y29v7yl5/qUdTb72EezuG6cpHmVZe6p0bFFzOIMgky+StCUmThvTawReKFiNoHH5GQQHgZVl11e4xzxtwPOB/xSRvcDLgNtr1WHs91yC3f1jrO1tOW6hte6WBEMTWQpF9e15602xqGTzxfmbhmIRW320TgxP5EodxYCtQBoCfgbBQ8B6EVkjIgngWuB270ZVHVbVHlVdraqrgQeAq1V1q49lKllagyAo7x8AWNSSoKjh+oOaLO1FUE0fgdUI6oFTI5gKgrZkDBHbt7iR+RYEqpoHPgDcAWwHvquqT4rIzSJytV/PW61l7SmOjWd9WQ00kytwcCg9Iwi82cVh2rs4nfN2J5u/aShfVPIFC4MgFYvKcDp3XNNQJCK0p2yZiUbm68xiVd0MbJ527BOznHu5n2WZzhs5dGQkw5ndLaf0sZ85Oo4qpTkEnu4WZwnsY+Ph+YOab79ij7ddZbZQJBYN5TzHujCSyVHUqXWGPLbMRGML7V+cnzuVVRo6CrCoxaluh6lGUG0QeLfbXIJgDU2bVeyxFUgbW2iDwM+9i3f3jSMCa3qsRpBeYI0gY3MJAuWtM9RZIQhsl7LGFfog8KtGsGJR04wPv3DWCObeuN7jDS+1GkGwvA/76U1D7U22S1kjC20QtCZjtKViHPZhUlmlEUPgdIi2JmOh2sR+suoagds0ZCOHAjV95VFPpzUNNbTQBgHMP5dAdeHj/YtFZU//eMUgAGfkUJg2sfeaeqptGrJlJoI1Xx/BifxNmPoX6iBY2tE063pD+UKRTZ/5KZ+5+6kFPeahkQzpXGHWIFjUkghVjSCddb7hV7NDGViNIGhDE1lEoC01MwjyRWU8a0HdiEIdBMvaZ68R3LntCDuPjPK9Rw4s6FvQ1PaUlYekdreEa5mJqVFD1kdwOhicyJX2zihny0w0tlAHwdKOFEfHJslW+Bb65Z89A8D+Y2medj/cq1Fpsbly1jRUmTUN1Yfp6wx5SiuQ2sihhhTqIDijM4Uq9I0eXyt44uAwD+0d5LdfsQaAu7cfqfoxd/eP0dEUp7tl5h8TOEEwMJ4NTVtrOmudxaeT4XRuxtBRmFpvaCgdni8xYRLqIFjaUXmnstt+9gwtiSgfvGI9L1jewd3bFhAEfeOs622ZdW/XrpYEk/kiEyFpa51aa2j+Hcqc88Px71KvBieydDbNDAKvRmBDSBtTqIOg0k5lfaMZ/u2x53jLS1bQnopzxbmLeXT/EEfHqhv7P9vQUc/UekPh+GaVyRUQgcQ8y0ZYH0F9GBzPVWwa8uYVWB9BYwp1EFSaXfyNB54lV1De+fLVAFx57hJU4b4dffM+3kgmR9/oJGvnCoLm8AVBKhadtYbkSVnTUF0YmsjOmEwGtktZowt1ELQlY7QkoqUawWS+wDd+vo9XPq+39GF+3hntLOtIVdVPsKd/HJh9xBBAV6sbBCHZ7Smdm393MpiqEfixGqypTjZfZDxbmDGHAKAlESUaEVtmokGFOghEhKUdKQ6POLOL//2xQxwdy/Jbl6457pxXnbOYnz59dN4PqdLQ0VlGDAGlTuSwbGKfyRVJxeZ/m3lNR1YjCI7XEVyps1hEbOG5BhbqIABnFdJDwxlUldt+9gzrF7dy6Vk9x51z5YYlTGQLPLBnYM7HumfHEVqTMVZ1Nc96zqIQ9hHMN2IIIBaNEIuIdRYHaLZ1hjy2zETjCn0QLO1IcXg4w0N7B3nyuRHedcnqGe3ZF6/tpjkRnbN56ImDw2z+5WF+69I1xOfoGG1LxohHJTRNQ9UGAXjbVVqNICje/JZKncXgDCG1IGhMoQ+CZR0p+kYn+eJP99DRFOdNF6yYcU4qHuUV63u4d3vfrOP///bOnXQ2x3nPK9ZUvN0jIs4m9mFqGppn6KgnGbcN7IPk7UBWqWkIbE+CRhb6IFjakaJQVO7cdoRrL1w5a8fmFecu4bnhDNsOjcy47aG9x/jPnf2877J1tKcq/xGV6wrRekMLrhFY01BghmbZi8Bju5Q1rtAHgTeXIBoRfuPi1bOe96pzFiMCd287fhipqvI3d+ykty3JO+e4f7nu1kRpud9Gl8kvNAisRhCUwdLKo5WbhqxG0LhCHwRL253ZxZvOW8ryzqZZz+tpTXLByk7u2XF8P8FPnz7Kg88c44OvOquqYZLg/KGFpbM4nS3Mu/KoJxmLWh9BgAYnsiSiEZpneR97QVAshmN5lDAJfRCctbiVN714OR++cv28515x7hIePzDMEXfpalXlb+/cyfLOJq596aqqnzNMK5BmcsXSHIH5JOPWNBSk4QlnnaHZJv91NMVRhdHJfI1LZvzmaxCIyCYR2Skiu0Tkxgq3f0REtonI4yJyj4ic6Wd5KknEInz6rS9i/ZK2ec+98twlANyz3WkeuuPJIzx+YJgPX7meRBVj5T1dLUmG0zlyhcb/9jtpTUOnjcGJ7Kz9A2DrDTUy34JARKLALcBVwAbgOhHZMO20R4GNqno+8C/AX/tVnlPh7CWtrOxq4p7tRygUlb+7cydre1t44wXLF/Q4Xe7exWHoJ1hI01DKRg0FanAiN+scArBlJhqZnzWCC4FdqrpHVbPAt4Fryk9Q1ftUdcK9+gAwc+xmHRERrjhnCf+96yjfeWg/T/eN8ZFXn01sngXVputqSQLOAl+NLpNfwPDRWMSWmAjQ0ES24vISHi8IbJmJxuNnECwH9pddP+Aem827gR/7WJ5T4tUbljCZL/LJf3uSDcvaed3zly34MbwVSAfGq1vR9HSVKxQpFLW0oNx8kjGrEQRpaKLyyqOeDtulrGHFgi4AgIi8HdgIXDbL7TcANwCsWlV9p6wfXrq6i7ZkjNHJPL/32rOJROZeVbOSsCxFPbVNpc0jqHeqytBErvRhX0lnky1F3aj8rBEcBFaWXV/hHjuOiFwJ/BFwtapW/Iqsqreq6kZV3djb2+tLYauViEV480tW8Mrn9fLK5y0+ocfwgqDRt6xMe0FQ5bDaZNyWmAjKRLZAtlCcu0Zgu5Q1LD9rBA8B60VkDU4AXAv8evkJInIB8E/AJlWdf8H/OvHJq887qft7IzMafXax96FezeqjYE1DQfIGLszVR5CKR0hEI1YjaEC+1QhUNQ98ALgD2A58V1WfFJGbReRq97S/AVqB/ycivxCR2/0qTz2JRyN0NMUbvkbgBV1bFctugDUNBWm+lUfBXYq6OW7DRxuQr30EqroZ2Dzt2CfKLl/p5/PXs+4QrDf04DPOst0vXtVZ1fnePAJVnXdHM3NqlYKgwn7F5WyZicYU+pnFQVkUgtnFW3YPsLa3hcXtqarOT8ajqEKuYEsY1Fqpaahl9hoBOEFgw0cbjwVBQLoaPAjyhSIP7R3k4rXdVd8nGfN2KbPmoVqbb+VRj9UIGpMFQUAafb2hJ54bYWwyz8tOKAisw7jWppqG5q4R2C5ljcmCICCLWpylqGfb6OZ0t2W30z+woCBw5xtYENTe4ESOlkR03jWzbJeyxmRBEJDulgS5gjbsSo5b9gywfnErvW3Jqu9TqhHYMhM1NzSRnXPEkKejKc5oJk/BlqJuKBYEAfEm7jTilpW5QpGte49x8brqawPgzCMAZ+lqU1uDE1kWtcw/zNdWIG1MFgQB6Wp1g6ABVyB9/MAwE9nCgjqKgdK+BdZZXHtD6bnXGfJ02npDDcmCIK+ABjEAABBQSURBVCDdLY1bI3hgj9M/cNFCg8A6iwMzNJErfdufy9QyExYEjcSCICClpqEGHDm0ZfcA5yxtK62pVC2vaciCoPYGJ7JV1QhsT4LGZEEQkO4GbRrK5ots3XdsQaOFPNZZHIxCURlO5+ZcZ8hjQdCYLAgC0pyIkYpHFlwj6B+d5Kkjoz6V6uQ9dmCITK644I5ioLSBjdUIams0k0N17nWGPLYnQWOyIAhQV3OCgQX0EeQLRd7+xZ/zhlt+xqHhtI8lO3Fbdg8gAhet6Vrwfa1pKBiDpQXnFlAjaLCabNhZEASoqzWxoH2Lv/ngs+w8MkomV+BTP9ruY8lO3JbdA5y7tL2qb5fT2RITwZhagnr+/7NkLEoqbktRNxoLggB1tSSrXoF0cDzL3935FBev7eZDV5zNjx4/xM92HfW5hAuTyRV45NnBE2oWgrIagc0jqKlq1xnydDYlLAgajAVBgLqa4xyrct/iv7/7KUYzOW66egPvvWwtq7qauen2J8nWUTPKL/YPMZkvLnj+gCdpfQSB8NYZqqZGALYCaSOyIAhQV0uSwfH5/6B2HB7h6w/s4/qLzuScpe2k4lFuev0GdvWN8dX79/pf0Cpt2T1AROClJ9A/ANY0FJSF9BGArUDaiCwIAtTVEmdsMj/nB5+qcvO/baMtFecjrz67dPyKc5fwqnMW85m7n+LISKYWxZ3Xlj0DnHdGR1UTkyoRERKxiC0xUWNDE1kiAu1V7iRnC881HguCAHW1OAuyzVUruOPJI9y/e4CPvPrsGZuG3PT6DeSKyv/eHHzHcSZX4BfPDp1w/4DHtqusrf3HJvjOQ/tZ29tKJFLdrnCdtl1lw7EgCJA383Zgln6CTK7An2/extlLWrn+olUzbj+zu4X3/Y+1/PAXz5WWdQjKI/sGyRZOvH/AYxvY187RsUne8aWfM5kv8vnrX1z1/Tqa4rbERIOxIAiQFwSzTSr70n8/w/5jaW56/XnEopX/q95/+Vks72ziph8+Sa4w8wNUVekfnSRf4bZqHRic4GsP7OO3vvIQ13/xAe7b2TdjH4UtewaIRuSE+wc8yVjERg3VwGgmx7u+/CCHRzLc9q6XcvaStqrv29EUZyJbqPh+M6cnXzevN3PzguDgYJpCUYmWVc0PD2e45b5dvPa8JVxyVs+sj9GUiPKJ12/gvV97mC/+9BkuXtfN9kMjpZ8dh0YZncwTjQgrFjWxqquZM7ubObOrhVXdzXQ0xUnFoyRjkdLvZCzCU0fG+M+dfdy7o4+n+8YAWNXVTKGo/OaXH+KFKzv58JXrufzsXkSELbsHeMHyDlqTJ/eWSsatachvmVyBG/75YXYcGuUL79zIS85ctKD7ly8z0dNa/X4Tpn5ZEASoty2JCNz4/V9y4/d/SXsqxqKWBJ3NCUYzOfJF5Y9et2Hex3nNhiVcdnYvf/UfO0rHWpMxzlnaxhsuWM6anhYGxifZNzDBvoEJHtv/HCOZ+TfEiUeFi9Z087aXruSV5yxmbU8LuYLyvUcO8Ll7d5UC4f2XreOxA0O8+9K1J/XvAf41DakqOw6Pcte2I9y7w6nRrCwLRe/y0vZU1W3llWRyBbbsGeD+XUc5o7OJl6/r4ewlrYic+GOeSoWi8r++8wu27Bng79/2Ql75vMULfozypagtCBqDr0EgIpuAfwCiwBdV9S+n3Z4E/hl4CTAAvE1V9/pZpnrS0RTn6+++iKePjDI4kWNoIsvgRI7BiSzFovLxq85hVXfzvI8jIvzVm8/n+48eYG1PKxuWtbNiUdOcH2hDE1mePTbBWCZPJl8gkysy6f7O5Aqc0dnEJWf1zPiGn4gJ1124ije/eEUpEN739YcBTrqjGLzO4vmDQFV5bjjDkweHefbYBG2pGN0tSbpaE3S3JOhuTZKKRXhw7zHu2naEu7Yd4cBgGhF44YpO2pvi/PLgMD9+4vBxu211tSS44pzFvPa8pVy6voeUu33mXI6OTXLvjj7u2X6Enz59lIlsgXhUyBWcx+1pTfLydd1cclY3L1/Xw8qu+f9P/aCq/MkPn+DHTxzmT351A2+8YMUJPU67txS1zSVoGOLXnrkiEgWeAl4NHAAeAq5T1W1l5/wOcL6qvk9ErgXeqKpvm+txN27cqFu3bvWlzGbhsvki33vkAI/sG+TP3vD8qj4453LtrVt4/MAwz1/eweK2JL1tSRa3pVjs1p62PTfCtkPOz3wfRCKgColYhFec1cOVG5ZwxbmLWdyWKp2TLxR5bijDs8cm2HdsnAefOca9O/oYzeRpTkS57OxeXnPeEjae2cXAeJbDwxmOjGQ45P7ec3Scxw8MoQpL21Ncce5irtywhIvXdnN0bJL7dw3ws91HuX/3AP2jzqCAtlSMpe0plrSnWNyeLF1uikcZdL8MDE1kOTaeZWgiRyZfYHFbimUdKZZ2TP1e0p6iNRkjFY/SFI8Sj0qp5pEvFDk8kuHgYJoD7s8Tzw1z17Yj/M7l6/j9Teec8P/RI88O8qbP388f/8q5nL+ik0JRKaqSd3+nYlHam2K0p+K0N8VpS8YWVMtSVSbzRSbzReJRIRGNEI1I3dSqPIWiMjaZZ3wyz9hkvvReS8Yix/2ORyLMVnRVKKqiuL/VOaYoRS0/5lxPxiK0nGDzq4g8rKobK97mYxBcDHxSVV/rXv84gKr+Rdk5d7jnbBGRGHAY6NU5CmVB0Nju2X6EHzx6kL7RSfpHJ+kbyTCeneozSMQinLO0jfPOaGfDGR2cd0Y7a3taGJvMc2w8y8B4lmNjWQbGJxlO53jB8k7+x9k9NCeq/+PJ5os8sGeAO7cd5s4nj9A3OnNUVzwqLG5LsbyziZef1c2V5y7hvDPaZ/2wUlV29Y1x/+4Bnjk6zuHhDIdHMvSNZOgbnSRfViuJR4XO5gSLmuN0NidIxiL0j05yeCQzZ/hFI0IqFiEZjzKczs3YV3hxW5I3XrCcG68656Q+VA8OpbnkL++t+nwRSoEVFSEiTi02EoGoCIrTpObVRivVCEUgEY04P7EIIlL6cBX3dmHmY0fKzvM+WAtFpVhUCu6Hq/Np437gQmkghHNfIeo+lve4mVyRsUyedADLpb/vsnXceNWJhXhQQfAWYJOqvse9/g7gIlX9QNk5T7jnHHCv73bPOTrtsW4AbgBYtWrVS/bt2+dLmU19Gp/MOx+WhSKre1qIzzKCyg/FovLYgSF2HB6ltzVZ+ibe3ZI4qb6E6c8xMJ4lkyuwqCVBSyI66wd1Olvg8EiGQ8Np+kYmGc/mSx+g6WyBdM756WpOsGJRE8sXNbFiUTPLOlInXVsr9+izgwync8QikdKHbsytkWRyBUbSeUYyOUbSOUYyeUYzOTK5AsWi82HsfSgX3c+flLuYXSoeJRl3LieiEfJFJZcvki0UyZb9nsq4479FO9+wpx7bu4xCJCJExfmAdy47H+wgbpBMBYr7yBSKTjAU3OBQhVQ8SmsySmsyTksySlsqRksyRkSEyXyBrFub8X7PNrpKlVJQeQFWHmjebSJT15+/vJ2XnHliI/PmCoLTorNYVW8FbgWnRhBwcUyNtSRjrDnJ0UgnKhIRLli1iAtWLWxkzUKfo7etuk7XpkSUNT0trOlp8a081fDz38PUnp9frQ4CK8uur3CPVTzHbRrqwOk0NsYYUyN+BsFDwHoRWSMiCeBa4PZp59wOvNO9/Bbg3rn6B4wxxpx6vtW3VTUvIh8A7sAZPnqbqj4pIjcDW1X1duBLwNdEZBdwDCcsjDHG1JCvDa+quhnYPO3YJ8ouZ4Bf87MMxhhj5mZrDRljTMhZEBhjTMhZEBhjTMhZEBhjTMj5NrPYLyLSD5zo1OIe4Oi8ZzWmsL52e93hYq97dmeqam+lG067IDgZIrJ1tinWjS6sr91ed7jY6z4x1jRkjDEhZ0FgjDEhF7YguDXoAgQorK/dXne42Os+AaHqIzDGGDNT2GoExhhjprEgMMaYkAtNEIjIJhHZKSK7ROTGoMvjFxG5TUT63N3fvGNdInKXiDzt/m64XUVEZKWI3Cci20TkSRH5kHu8oV+7iKRE5EERecx93X/qHl8jIj933+/fcZeCbzgiEhWRR0Xk393rDf+6RWSviPxSRH4hIlvdYyf1Pg9FEIhIFLgFuArYAFwnIhuCLZVvvgJsmnbsRuAeVV0P3ONebzR54KOqugF4GfC77v9xo7/2SeBVqvpC4EXAJhF5GfBXwN+r6lnAIPDuAMvopw8B28uuh+V1v1JVX1Q2d+Ck3uehCALgQmCXqu5R1SzwbeCagMvkC1X9Cc7eDuWuAb7qXv4q8IaaFqoGVPWQqj7iXh7F+XBYToO/dnWMuVfj7o8CrwL+xT3ecK8bQERWAL8CfNG9LoTgdc/ipN7nYQmC5cD+susH3GNhsURVD7mXDwNLgiyM30RkNXAB8HNC8Nrd5pFfAH3AXcBuYEhV8+4pjfp+/wzw+4C3O3w34XjdCtwpIg+LyA3usZN6n58Wm9ebU0dVVUQadsywiLQC3wM+rKojzpdER6O+dlUtAC8SkU7gB8A5ARfJdyLyq0Cfqj4sIpcHXZ4au1RVD4rIYuAuEdlRfuOJvM/DUiM4CKwsu77CPRYWR0RkGYD7uy/g8vhCROI4IfANVf2+ezgUrx1AVYeA+4CLgU4R8b7oNeL7/RLgahHZi9PU+yrgH2j8142qHnR/9+EE/4Wc5Ps8LEHwELDeHVGQwNkb+faAy1RLtwPvdC+/E/hhgGXxhds+/CVgu6p+uuymhn7tItLr1gQQkSbg1Tj9I/cBb3FPa7jXraofV9UVqroa5+/5XlW9ngZ/3SLSIiJt3mXgNcATnOT7PDQzi0XkdThtilHgNlX984CL5AsR+RZwOc6ytEeAm4B/Bb4LrMJZwvutqjq9Q/m0JiKXAj8FfslUm/Ef4vQTNOxrF5HzcToHozhf7L6rqjeLyFqcb8pdwKPA21V1MriS+sdtGvo9Vf3VRn/d7uv7gXs1BnxTVf9cRLo5ifd5aILAGGNMZWFpGjLGGDMLCwJjjAk5CwJjjAk5CwJjjAk5CwJjjAk5CwJjphGRgruyo/dzyhaqE5HV5SvDGlMPbIkJY2ZKq+qLgi6EMbViNQJjquSuA//X7lrwD4rIWe7x1SJyr4g8LiL3iMgq9/gSEfmBu1fAYyLycvehoiLyBXf/gDvdGcHGBMaCwJiZmqY1Db2t7LZhVX0B8DmcmeoA/wf4qqqeD3wD+Kx7/LPAf7l7BbwYeNI9vh64RVXPA4aAN/v8eoyZk80sNmYaERlT1dYKx/fibAKzx13g7rCqdovIUWCZqubc44dUtUdE+oEV5UscuEtk3+VuIIKI/AEQV9VP+f/KjKnMagTGLIzOcnkhyte+KWB9dSZgFgTGLMzbyn5vcS/fj7MCJsD1OIvfgbNl4PuhtHlMR60KacxC2DcRY2Zqcnf88vyHqnpDSBeJyOM43+qvc499EPiyiHwM6Ad+0z3+IeBWEXk3zjf/9wOHMKbOWB+BMVVy+wg2qurRoMtizKlkTUPGGBNyViMwxpiQsxqBMcaEnAWBMcaEnAWBMcaEnAWBMcaEnAWBMcaE3P8HAJ4wawps66EAAAAASUVORK5CYII=\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model = Transformer(50).cuda()\n",
        "train_network(model, train_data, valid_data, batch_size = 512, num_epochs=50, learning_rate=2e-3)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "7TAt0y4sJH1J",
        "outputId": "7c695fa7-e0bc-4daa-9b90-92d380d22798"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1; Loss 0.721622; Train Acc 0.508922; Val Acc 0.506140\n",
            "Epoch 2; Loss 0.690972; Train Acc 0.508922; Val Acc 0.506140\n",
            "Epoch 3; Loss 0.687290; Train Acc 0.509101; Val Acc 0.506140\n",
            "Epoch 4; Loss 0.666188; Train Acc 0.774625; Val Acc 0.781100\n",
            "Epoch 5; Loss 0.439670; Train Acc 0.816203; Val Acc 0.811532\n",
            "Epoch 6; Loss 0.344043; Train Acc 0.863490; Val Acc 0.850507\n",
            "Epoch 7; Loss 0.294488; Train Acc 0.883655; Val Acc 0.864389\n",
            "Epoch 8; Loss 0.296965; Train Acc 0.906674; Val Acc 0.882007\n",
            "Epoch 9; Loss 0.263409; Train Acc 0.918273; Val Acc 0.886279\n",
            "Epoch 10; Loss 0.234826; Train Acc 0.916131; Val Acc 0.884143\n",
            "Epoch 11; Loss 0.230537; Train Acc 0.931299; Val Acc 0.890016\n",
            "Epoch 12; Loss 0.216208; Train Acc 0.930228; Val Acc 0.884677\n",
            "Epoch 13; Loss 0.181262; Train Acc 0.950214; Val Acc 0.900694\n",
            "Epoch 14; Loss 0.192971; Train Acc 0.952355; Val Acc 0.891618\n",
            "Epoch 15; Loss 0.148490; Train Acc 0.968415; Val Acc 0.903364\n",
            "Epoch 16; Loss 0.110398; Train Acc 0.974304; Val Acc 0.901762\n",
            "Epoch 17; Loss 0.109447; Train Acc 0.975196; Val Acc 0.896423\n",
            "Epoch 18; Loss 0.080498; Train Acc 0.981085; Val Acc 0.898558\n",
            "Epoch 19; Loss 0.087762; Train Acc 0.962884; Val Acc 0.891618\n",
            "Epoch 20; Loss 0.102617; Train Acc 0.975018; Val Acc 0.894287\n",
            "Epoch 21; Loss 0.077259; Train Acc 0.984832; Val Acc 0.896423\n",
            "Epoch 22; Loss 0.070493; Train Acc 0.987330; Val Acc 0.897491\n",
            "Epoch 23; Loss 0.099381; Train Acc 0.990721; Val Acc 0.903364\n",
            "Epoch 24; Loss 0.039832; Train Acc 0.981799; Val Acc 0.890550\n",
            "Epoch 25; Loss 0.086890; Train Acc 0.988401; Val Acc 0.897491\n",
            "Epoch 26; Loss 0.042337; Train Acc 0.992862; Val Acc 0.895889\n",
            "Epoch 27; Loss 0.075680; Train Acc 0.990007; Val Acc 0.896957\n",
            "Epoch 28; Loss 0.051325; Train Acc 0.992148; Val Acc 0.899626\n",
            "Epoch 29; Loss 0.046897; Train Acc 0.991792; Val Acc 0.895889\n",
            "Epoch 30; Loss 0.044702; Train Acc 0.986795; Val Acc 0.882007\n",
            "Epoch 31; Loss 0.031232; Train Acc 0.992862; Val Acc 0.896423\n",
            "Epoch 32; Loss 0.037260; Train Acc 0.995539; Val Acc 0.892152\n",
            "Epoch 33; Loss 0.022632; Train Acc 0.992327; Val Acc 0.888948\n",
            "Epoch 34; Loss 0.057032; Train Acc 0.990364; Val Acc 0.891618\n",
            "Epoch 35; Loss 0.016872; Train Acc 0.993219; Val Acc 0.904965\n",
            "Epoch 36; Loss 0.056667; Train Acc 0.994647; Val Acc 0.900160\n",
            "Epoch 37; Loss 0.031509; Train Acc 0.995004; Val Acc 0.901228\n",
            "Epoch 38; Loss 0.035260; Train Acc 0.994825; Val Acc 0.899092\n",
            "Epoch 39; Loss 0.018462; Train Acc 0.996431; Val Acc 0.895889\n",
            "Epoch 40; Loss 0.005488; Train Acc 0.996074; Val Acc 0.899092\n",
            "Epoch 41; Loss 0.036477; Train Acc 0.994825; Val Acc 0.895355\n",
            "Epoch 42; Loss 0.042941; Train Acc 0.994111; Val Acc 0.902830\n",
            "Epoch 43; Loss 0.039425; Train Acc 0.993754; Val Acc 0.902296\n",
            "Epoch 44; Loss 0.041870; Train Acc 0.996788; Val Acc 0.897491\n",
            "Epoch 45; Loss 0.030002; Train Acc 0.996253; Val Acc 0.899626\n",
            "Epoch 46; Loss 0.027623; Train Acc 0.995182; Val Acc 0.900160\n",
            "Epoch 47; Loss 0.028973; Train Acc 0.993398; Val Acc 0.903364\n",
            "Epoch 48; Loss 0.019620; Train Acc 0.992505; Val Acc 0.899092\n",
            "Epoch 49; Loss 0.038391; Train Acc 0.994111; Val Acc 0.898025\n",
            "Epoch 50; Loss 0.027236; Train Acc 0.994468; Val Acc 0.905499\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXxU9bn48c+Tyb6xJGFLgODGorKLa1VcURTqDtUqba/brbZ2sdX+rLu3m72tVq8t1qrXqrh7UUGriFtxISyyo8gaIBASsk6Wyczz++M7IQGyTCCTCTPP+/WaV2bOnDnznMnMeb7LOd+vqCrGGGNiV1ykAzDGGBNZlgiMMSbGWSIwxpgYZ4nAGGNinCUCY4yJcZYIjDEmxlkiMDFDROaKyDWdva4xhzqx6whMdyYiVc0epgJ1gD/4+HpVfbbrozo4IpIJ3AtcDPQGdgBvAPer6q5IxmZik9UITLemqumNN2AzcGGzZXuSgIjERy7K0IlIIjAPOBqYBGQCJwIlwIQD2N4hsd+me7NEYA5JInK6iBSKyC9FpAh4UkR6icibIlIsIruD9/OaveYDEfmP4P0ZIvKJiDwYXHeDiJx3gOsOEZGPRKRSRN4TkUdF5J+thH41MAi4SFVXqWpAVXeq6n2qOie4PRWRI5pt/ykRub+N/V4tIhc0Wz8++BmMDT4+QUQWiEiZiHwpIqcf7OdvooslAnMo64drWhkMXIf7Pj8ZfDwIqAEeaeP1xwNrgWzg98ATIiIHsO5zwBdAFnA38N023vMs4G1VrWpjnfbsu9/PA9ObPX8usEtVF4tILvAWcH/wNT8HXhGRnIN4fxNlLBGYQ1kAuEtV61S1RlVLVPUVVfWqaiXwAHBaG6/fpKqPq6ofeBroD/TtyLoiMgg4DrhTVetV9RNgdhvvmQVs79hu7mev/cYloikikhp8/ju45ABwFTBHVecEax/vAgXA+QcZg4kilgjMoaxYVWsbH4hIqoj8TUQ2iUgF8BHQU0Q8rby+qPGOqnqDd9M7uO4AoLTZMoAtbcRcgksiB2Ov/VbVdcBq4MJgMpiCSw7gag2XBZuFykSkDDilE2IwUcQ6msyhbN9T3n4GDAWOV9UiERkNLAFaa+7pDNuB3iKS2iwZDGxj/feA+0UkTVWrW1nHiztDqlE/oLDZ45ZO9WtsHooDVgWTA7ik9IyqXtvOfpgYZjUCE00ycP0CZSLSG7gr3G+oqptwTS13i0iiiJwIXNjGS57BHZxfEZFhIhInIlki8isRaWyuWQp8R0Q8IjKJtpu3Gs0CzgFupKk2APBPXE3h3OD2koMdznktbsXEJEsEJpr8GUgBdgGfAW930fteSdMpoPcDL+Cud9iPqtbhOozXAO8CFbiO5mzg8+BqP8Ylk7Lgtl9vLwBV3Q58CpwUfP/G5VuAqcCvgGJcEroV++2bZuyCMmM6mYi8AKxR1bDXSIzpDFYqMOYgichxInJ4sJlnEq4E3m4p3pjuwjqLjTl4/YBXcaeGFgI3quqSyIZkTOisacgYY2KcNQ0ZY0yMO+SahrKzszU/Pz/SYRhjzCFl0aJFu1S1xaFFDrlEkJ+fT0FBQaTDMMaYQ4qIbGrtOWsaMsaYGGeJwBhjYpwlAmOMiXGWCIwxJsaFLRGIyD9EZKeIrGjleRGRh0VknYgsa5xNyRhjTNcKZ43gKdycrK05DzgyeLsOeCyMsRhjjGlF2BKBqn4ElLaxylTgf9X5DDeBiE2WYYwxXSyS1xHksvdMToXBZftN4yci1+FqDQwaNKhLgjPGdC1Vpczro6iilqKKWsq9PjxxQoInjsR499fdj6NnSgJZaUlkpsTT+jTTHRMIKBtKqlmxtZyNu7zEe4TkBA/JCXEkx3tITvCQkhhHZnICPVMTyExJoGdKIonxh35X6yFxQZmqzgRmAowfP94GRzKmHcWVdXy1o5I1RZV8VVTJ2h2VfL2jksyUBEbm9WDUwJ6MyuvJsXk9yExO6PD2y7z17Kyso7jZbVdVHcVVddT6/Pj8is8fcLcGpd4fQAGPgCdOiBPBE+dutT4/RRW17Kioo74h0KE44uOEnqmJZKUl0jstkYzkeNKS4klN9DT9TYwnKSFuTyJJ8AiJwfvlNT5WbCtn5dYKVm4rp7re3+HPIjXRQ8+UBHIykxnQI5kBPVPcLXg/OyOJjOR40hPjiYvbP2nVNfjZWVHHjmAC3FVZR70/gM+v1DcEmj5HvzJ5ZH+Oy+/d4RjbE8lEsJW9p/TLCy4zJiL8ASVOOOASZml1PR99Vcz8tTsp2LibvplJjBiQyfD+mYzon8nQfhmkJrqfXK3PT+FuLxt3edlU6mVTSTVVdQ17DlAJnjgS4t0BKzs9iTOH9yGvV2qb779hVzWzFm7m9SVb2VHRNC9O77REhvbN4JJxeZR5fXxZWMY7K3fsef6wnDSG98skr3cKA3ulktcrhYG9U8ntmUKCJ44Nu6pYua2C1dsrWb29glXbKyiu3H/encT4OLLTEklNincld09TKT4jwZXcAwHFH1D86g5yflUSPXGMHdSLfpnJ9A3e+vVIoldqIgFV6huakkq9P0CdL0BZTT0lVfWUVtez29t0f3OpF2+9H299A9V1fmp87R/YkxPiGNE/k0vH5XF0bg+Oze3B4TnpBFSp8wWobfBT6/NT6wtQ4/NTWeujzOujrMZHubd+z/0dFbV8taOSD9YWt/q+6UnxZCS7W5wIOypq2e31tRlfQrPPccSAzLAkgrCOPioi+cCbqnpMC89NBm4CzgeOBx5W1QntbXP8+PFqQ0yYzlBUXsuSzbtZsqWMxZt2s3xrOQN7p3LH5OGcPrRPu69XVVZsrWD+2p3MX7uTpVvKUIXs9ESOH5JFcWUdq7dXUFnXAIAI5GelUefzs72iluY/vYykeDJTEmgIuJKfr8Ed9Or9gT3rHZvbg0nH9OO8Y/pxWE464BLK2yuKeP6LzXy+oRRPnHDGsD6ccFgWw/plcFTfDHIykvaLvcxbz7LCcpYVlrF0SznfFFexdXcN9f69S+SJnrg9yxI8whF9MhjRP5Nh/TLo1yOZnIwkcjKSyE5PIjO585ppOksgoHh9fuqa1VLqm9VUUhLjGJKdjqeFkvqBUlXKa3xsLathW1ktpdV1VNY2UFHbQGWtj8rgX38A+mYmNSXAHsn0y0wmOz2R5ATPntpLZ32mIrJIVce3+Fy4EoGIPA+cjpuCbwdu/tgEAFX9q7i9ewR3ZpEX+J6qtnuEt0RgOqrW52d9cTVf76zkm51VfLWjii8Ly9heXgu4g90xuZmMzOvJB2t3srHEyxnD+nDH5OF7DrjN1dT7eX3pVp7690bW7qhEBEbm9WTi0BwmDu3Dsbk99jQBqCqFu2tYtb2CVdsqWFtUSWqih0FZqeRnpTE4K5XBWWn0Sk1o9Qe/cVc1b68sYu6KIr7cUgbA0L4ZHJ2bybzVOymv8TGodypXHDeQy8bl0Scz+YA+p0BA2VlZx5bdXgp3e9lSWkN1XQNH9c1geP9MjuiTHhXt4bEqIokgXCwRmFCs2FrOX97/mjVFlWwu9e4pVccJDM5K49jcHowZ1JMxg3oxvH8GSfEewLXXPr1gIw/PW0ddg5/vnTyEm844gszkBLaUevnnZ5uYtXAL5TU+hvfP5OoTB3P2iL5kp+9f6g6HbWU1vL2iiLdXFLFiWzlnDOvD9AmDOPGwrBbbn41pZInAHDJKq+tJTXRnaBwIb30Df3r3K574ZAO9UhM54fAsjshJ58i+6RzZJ4P87NQ9B/22FFfW8Yd31vDSokKy0hL31BZEhHOP7suMk4ZwXH6vbtcUYkxrLBGYbq2+IcC81Tt4aVEhH6zdyaiBPXn+2hM6nAw++qqY//f6craU1jB9wkBumzScHqkdPyOmuWWFZdz35io27PJy+fg8rjphMAN6phzUNo2JBEsEpltavb2ClwoKeX3pVkqr6+mbmcRpR+XwYkEhF4zsz1+mjwmpxF1aXc99b67itSVbOSw7jf+6+FhOOCyrC/bAmENHW4ngkLiOwByaNu6q5v01O9ntrQ+eNdF0xkRxZR3fFFeT4BHOGdGPS8fnceqROXjihCHZ6fzu7TUclpPOT88+qs33+NfKIm57dTkVNT5uPuMIfjjxiANuVjImVlkiMJ3KW9/AnOVFvFiwhS82uBFG4qTx/OmEPedQ52elcdUJg5k6OpfeaYl7beOG0w5jw64qHp73NYdlp/HtMbn7vY8/oPz3u2t5dP43HJObyfPXnsDQfhldso/GRBtLBOagqSqLN+/mpYJC3vhyG9X1fvKzUrn13KFcNCaX/j2SO9SpKiLc/+1j2Vzq5RcvLyOvVwrjm11Es7u6nh/NWsLHX+9i2nEDuXvK0VYLMOYgWB+BOSiBgPKr15Yza+EWUhM9TD62P5cfN5Dxgw/+jJoybz0X/c8Cymt8vP6fJzMoK5UVW8u54Z+L2FlRxz1Tj2b6BBt7yphQWB+BCQtV5e43VjJr4RauP/Uwbj7zSNKTOu8r1TM1kX/MOI5vP/pvvv/0Qq45KZ/731xF77REXrzhREYP7Nlp72VMLLPLBM0BUVV+M3cN//vpJq4/9TBuO29YpyaBRkOy0/jrVePYVFLNr19fwdhBvXjj5lMsCRjTiaxGYA7In9/7mpkfrefqEwdz23nDwnph1YmHZ/E/V45j3c4qrv3WEOI9Vn4xpjNZIjAd9tcPv+GheV9z2bg87r7w6C65uvbsEX05e0TfsL+PMbHIilamQ55esJHfzl3DhaMG8NtLRtr4NsZEAasRmFb5A8r28ho2lXjZWFLN6u0V/POzzZwzoi//ffmoTh261xgTOZYIzF78AeW+N1fx0dfFFJbuPT59YnwcF4zszx8vH0WCtdMbEzUsEZi9/OX9r3lqwUbOGNaHc0b0Iz8rdc/Y+f0yk60pyJgoZInA7PHB2p08NO9rLhmbx4OXjbQhlo2JEVa/NwAU7vZyywtLGdo3g/u/fYwlAWNiiCUCQ12Dnx8+uxi/X3nsqnGkJNq4PcbEEmsaMtz/5mq+LCznr1eNY0h2WqTDMcZ0MasRxLjXlhTyzGdumIhJx/SLdDjGmAiwRBADSqvrqaprYN+RZtcWVXL7q8uZMKQ3t547NELRGWMizZqGopiq8vt31vLYB98AEB8n9ExNoEdKAj1TE9lS6iUjOYFHpo+x8XuMiWGWCKKUqvLAW6v5+ycbuGhMLsP7Z1Dm9VFW46Pc66Ospp787DR+OWkYfTKTIx2uMSaCLBFEIVXlnjdW8dSCjcw4KZ+7Lhxhp4MaY1pliSDKBALKXbNX8sxnm/jBKUO4Y/JwSwLGmDZZIogigYByx/+t4LnPN++ZLMaSgDGmPZYIDiFl3nrumr2Sr3ZUMbh3KoOz3RhAg3u78YD+Mm8dLxRs4T9PP5xbzx1qScAYExJLBIeIxZt3c/NzS9hZWcuJh2fz9c5K3l+zc6/RQQF+dMYR/OTsoywJGGNCZomgm1NVnvhkA7+du4b+PZN55caTGJnn5uttnC9gc4mXjSVeeqclMOmY/hGO2BhzqLFE0I2Ve338/OUveXfVDs49ui+/v3QUPVIS9jzviRPyeqWS1yuVk46IYKDGmEOaJYJuaumWMn747GJ2VtZy5wUj+N7J+dbcY4wJi7BeTioik0RkrYisE5HbWnh+sIjME5FlIvKBiOSFM55DxcZd1Uyf+Rki8NINJ/H9U4ZYEjDdR81u+OJx2PQpBPyRjib6qcKGj+DpC2HTgrC8RdhqBCLiAR4FzgYKgYUiMltVVzVb7UHgf1X1aRE5A/gN8N1wxXQoCASUW1/+kgSP8PINJ9GvRze/6jcQgE/+CMtfgbRsyOgPGf0gc4C733Mg9B8Ncd14aOv6alj1f7D8Jcg6As6+DxK6+eceKbs3wrOXwa6v3OPUbBh6HgybDIedDgkpLb+ucZyrri7QVJfAwsehdD3EJ0F8ivvfNv7NGABHng2pvTvn/QIBKFkH2xbD1sWwbQnExcPhE+HwM2FAB34LqrDuPfjoD7Dlc0jvC96SzolzH+FsGpoArFPV9QAiMguYCjRPBCOAnwbvzwdeD2M8h4QnF2xk4cbd/PGyUZFJAr4aePOn0CMXTvkJJLYxLHX1Lnj1WvjmfRh0Ivh9sOUzqCwCf33Teun94OiL4JhLIG981x8MWqLqflxL/gkrX4P6KugxyO3LtiVwxbOQ0bdr4igvhJ2rIT4R8o5r+zOPpMICeH6a+z9f+TLUVcCat1wSXfIMJKTCYRMhKd3VGmp2Q02Z+1tb5var1xDoPQR6Hxa8fxhkHwXpOaHFEAhAVZH7TsW10aBRWQQL/gIF/wCfF3oOgoY68NVCQ83e30+Jg0EnuWQ27Hzolb/3tlTdAbh0A5Rtct+Vxu3s+VsDxWtg21L3uQAkpLkDv88L8/8L5j8AKb1cwjz8TPebScuGpMy99yUQgLVzXALYvhQy8+D8B2HMd8NWQJF9R6TstA2LXApMUtX/CD7+LnC8qt7UbJ3ngM9V9SERuRh4BchW1ZJ9tnUdcB3AoEGDxm3atCksMUfa+uIqzn/4Y045IpvHrx7f9c1BDXUw6zuuFAKutHTOfe4Avm8smz+Hl2a4H8h5v4NxM5rWUQVvKVRug+K17kD79bvgr3M/yGMugRHfduuWrofdG9zf0o2uxJl9JJx4Exxx5oEnDVV38PGWNh2MGm9VRe7gVbLO/ViPuQhGXwWDToDVs+G1GyClN0x/HvqPbP09ipa7zyqld7AG1M99Zqm9m+IOBNyBofFg6C2Fkm9g56rgbXXTgQNc6bH/aBh8Egw+2cWU0rNj+x4ItH2QPBCrZrukn9HPJYHsI5uea6iHjR+7pND43Unp5eJO6eVuyT3dfpYG/9dlm0GbNSv1GAgDxkDuWMgd5z6DpAyXJLcualbCXgr1lZDcwx28Gz+n/iPBk+DW//dDsOhpCPjgmEvhWz+FPsP3/4waamHXWlgzx8W+c6V7ru+xbrtVO4Lfz417/4/25UlyB+jeh8GAsW4fBoyFnKFNpf/qElg/3xU01s1z38FGEuc+n8bPq7bcfTd7DYFv/QxGXuEKCQdJRBap6vgWn4twIhgAPAIMAT4CLgGOUdWy1rY7fvx4LSgoCEvMkeQPKJf/7VPW7azi3Z+c2vUDwfl97sC+5k248CHIGQZzboWiZe6Hdt7vod8x7gD76aPw3l3QIw8ue9qVetpTW+5+bMtfhvUf7H0QAEjr435IPQfCxk+gcjvkDIcTfwgjL3fV+pZUFbsfc+n6poPM7g0uqdSVtx7PoJNgzFUwYqorwTa3/Ut4fro7eF/0Nxgxpem5xur6gr/Ahg9b3rYnEdJyXEmwthw0sP86yT2h79HuANVnhPvr87o24E0L3MHPXw8I5J8ClzwRWg1lzRx47XqXjA4/0yXTIae6g+qBUIXP/gfe+X+uNjd9livFHiy/D8q3uP/XztXBg/xid9AFQNzBvjZ4KIhLcN+/AWNdDWLnSvc5laxzzyekQb9j3eeGwqjprkabdXjoMZWub0oK25a45L5v7aXXYFeCT0iB+GR362jSVXWFgO1f7l9zqtkNgQZX+j/6IvB0XqNNpBLBicDdqnpu8PHtAKr6m1bWTwfWqGqbHcbRmgj+/vF67n9rNX+6YhQXjeniPvOA35X2VrziDvjHX9+0fPH/wrx73Q9y/A/cAXrNmzDsApj6aMdLq+CalL5+1zUV9B7ifmTND8YN9bDyVXew3bHCJYnjr4P8b7kaRmNpescq8O5qel1cvKtxNP5we+W7g1ZjqXRP6bSHKz22pbIIZl0JWwvgjDvgxJth+YsuCRavcSX/46+H0Ve6poHKIqjY5v5WbnMJKjFt71JxY8m4V74rWbdV2/HVuIPaho/d55CWDVe9CtltnCf85Qvw+o0uwWQOcK/1VbvPZeDxcPgZTaXnUJqffLXw7q/hi5kwfApcPLP1PoDOUl3iDsLbFkPFVuh7jCth9z2m5cJA5Q7YvAA2/tv9r3LHwck/dt8Ds5dIJYJ44CvgTGArsBD4jqqubLZONlCqqgEReQDwq+qdbW03GhPBN8VVnP/Qx5x6VA4zvzuu85qEVF0b45YvXLtkTguTzwQCMPsmWPosnHUPnHLL/ut4S10bZ8ETrhp71j2upB7upitVV3v49JGmJgdwbdF9hjeVpnOGuYN/j4GdWoLCVwuzgwkgIc0dVPseCyfdBEdf3CnV9ZAULoLnLnP3r3zJHez29flMmHurS5bTn3c1gIZ61w/yzTzXHFG0zK0rca62lTsm2JQxziWGxqaqxiRb+o2rzZx0M5x1b+c3N5kuFZFEEHzj84E/Ax7gH6r6gIjcCxSo6uxg89FvAMU1Df1QVeva2ma0JQJ/QLn0rwvYsKuaf/3kVPpkdEKTUPUuWPai6wjdubJped9j4dhLXBt9z0HuQPvWz9wB/rTbYOLtbW+3eK2r0vc75uBj7Kida1yTT84w6Dm46w5Kqq5EvuVzmHAtDDktMp3dJd/AMxdBdTFc/gwceVZTfB8/CO/fD0Mnw6X/aL1DsarYlZobm2G2LnJNEXsRV0trbK4adKJrXjKHvIglgnCItkTwtw+/4Tdz1/DQtNFMHZ3b8ko1u2H9h65k980H7qyFxiaV3oc1tWPWlMHSf8Lat11H2YCxMOZKd/Ba955r+ilc6LaZN8E1H6x63VWlz7qne5zNY1pXuQOevcSV2qc+6joR/3WHqzGNnOaWdaRGpOra5LctdrWfPsNdok1MDdsumMixRNBNrSmqYMoj/2bi0Bz+etU+TUI7VrkzWNbNc6U4DbhOqiGnuvbi0g2uhFxeuHdnZGo2jJrm2q77jtj/TXdvdAlhxauu/X3C9e6sH0sCh4baCnjhSneB0cAT3Om6E66DSb+zphvTJksE3VBlrY8pj/yb6roG3vrRt8jJaNYRtv4D+OclrrM2d2zT2R+54/bv5Gyod6fi7d7gHg85LfS268odkN7HksChpqHOneK68lU49Rcw8Vf2PzTtaisR2FhDEaCq3PbKcjaXennuP47fOwnsWAUvfNedInf1/7kDdVviE92ZJG2dTdKarrhgynS++CR3OumZv3ZNgsYcJKtLRsBTCzby1vLt/OLcoRx/WFbTExXb4NlL3RkcV77UfhIwsSsuzpKA6TRWI+hiizbt5oG3VnPW8L5cd2qzH3JdJTx7ubsA6Xtz3cVaxhjTBSwRdKHS6npuem4x/Xsm88fLRjV1Dvt98OI17vztK19qe1gDY4zpZJYIukggoNzywlJKqut59caT6JEa7PRVhTd/4k4NnfKInbNtjOly1kfQRR6Zv46Pvirm7guP5pjcHk1PfPSgG7nxtF/C2JgegdsYEyFWI+gCizaV8qf3vuLiMblMnzDQLVSFT/4E8+93A2Sd3s5VvcYYEyaWCLrA819sIT0pnvsvOsb1CwT88Pbt8MXf4NjL4cKH7TxwY0zEWCIIM58/wLurdnD28L6kJsa7i4Fevc4N7XDiTW42LLsi1BgTQZYIwuzTb0oor/Fx3rH93amhs650k3icc78b1dEYYyLMEkGYzV2xnbRED9/q1wBPTobi1XDx426yFWOM6QYsEYRRgz/Av1bu4JIjIPnpSW5ax++8aKeIGmO6FUsEYfTFxlJKquv5AW9A9U53xXDu2EiHZYwxe7FeyjB6e0URvRLqGVT4hpus3ZKAMaYbskQQJoGA8vaKIm7ptxKpr4Jx10Q6JGOMaZElgjBZvHk3OyvruND/rhtSetCJkQ7JGGNaZIkgTOYsL+Joz1Z6ly6FsVfbBWPGmG7LEkEYqCpvr9jOT3ovAE8ijPpOpEMyxphWWSIIgy8Lyykpr+DU2nkw7AJIy2r/RcYYEyGWCMJg7ortTI5fSKKvwjqJjTHdnl1H0MlUlbnLi5iZ9hGkDoH8UyMdkjHGtMlqBJ1s1fYK4nevY1jdctdJbAPKGWO6OTtKdbK3VxQxzTMfjYuH0VdGOhxjjGmXNQ11sneXbebFxE+QoedBRt9Ih2OMMe2yGkEn+npHJYeXfkRmoBzGzoh0OMYYExJLBJ1ozvIipse/jz8jDw6fGOlwjDEmJJYIOomq8vniRZwStwLPuKshzhPpkIwxJiSWCDrJ4s1lfKviTQLEwZirIh2OMcaEzBJBJ5nzxUqu9ryLf9iF0CM30uEYY0zIwpoIRGSSiKwVkXUiclsLzw8SkfkiskRElonI+eGMJ1xqfX76r3ycFKkj4YzbIx2OMcZ0SNgSgYh4gEeB84ARwHQRGbHPancAL6rqGGAa8D/hiiecPlq6muk6l12DJ0Of4ZEOxxhjOiScNYIJwDpVXa+q9cAsYOo+6yiQGbzfA9gWxnjCpuHjh0gWH1mT74x0KMYY02HtJgIRuVBEDiRh5AJbmj0uDC5r7m7gKhEpBOYANx/A+0RUcdFmJpa/zursc/D0GRrpcIwxpsNCOcBfAXwtIr8XkWGd/P7TgadUNQ84H3impaQjIteJSIGIFBQXF3dyCAen6K3fkoiPjHN+FelQjDHmgLSbCFT1KmAM8A3wlIh8GjwwZ7Tz0q3AwGaP84LLmvsB8GLwfT4FkoHsFmKYqarjVXV8Tk5OeyF3Ga3YzlFbXuKD5DMYfNSoSIdjjDEHJKQmH1WtAF7GtfP3By4CFotIW005C4EjRWSIiCTiOoNn77POZuBMABEZjksE3avI34aSd36HRxson/CTSIdijDEHLJQ+giki8hrwAZAATFDV84BRwM9ae52qNgA3Ae8Aq3FnB60UkXtFZEpwtZ8B14rIl8DzwAxV1YPZoS5TsY2eq57lNT2ViSccH+lojDHmgIUy+uglwJ9U9aPmC1XVKyI/aOuFqjoH1wncfNmdze6vAk4OPdzuw//RH0H9LM2/lsvSEiMdjjHGHLBQEsHdwPbGByKSAvRV1Y2qOi9cgXVrZVuQxU/zQsNpnH7CcZGOxhhjDkoofQQvAYFmj/3BZbHrk/8mEFCeTbyM04d2n85rY4w5EKEkgvjgBWEABO/HdFuIf/1HfBAYzYQxo0jw2HBNxphDWyhHseJmnbuIyFRgV/hC6v7qqsrYGcjgkrF5kQ7FGGMOWih9BDcAz4rII4Dgrha+OqxRdTqmw+IAABYwSURBVHNxvkoSU3tw9IDM9lc2xphurt1EoKrfACeISHrwcVXYo+rO/D6StY7UzN6ISKSjMcaYgxbS5PUiMhk4GkhuPPip6r1hjKv7qqt0f5Pau7DaGGMODaFcUPZX3HhDN+Oahi4DBoc5rm5L6yoAkGRLBMaY6BBKZ/FJqno1sFtV7wFOBI4Kb1jdV21VGQBxKT0iHIkxxnSOUBJBbfCvV0QGAD7ceEMxqabSJQJPsiUCY0x0CKWP4A0R6Qn8AViMm0zm8bBG1Y011ggS0uyMIWNMdGgzEQTnBpinqmXAKyLyJpCsquVdEl03VFftEkFSWq8IR2KMMZ2jzaYhVQ3g5h1ufFwXy0kAwOd1u5+S3jPCkRhjTOcIpY9gnohcInbSPAD+mmAiyOgd4UiMMaZzhJIIrscNMlcnIhUiUikiFWGOq9sK1FTQoHGkp6dHOhRjjOkUoVxZbCfMN1dXQRUpZKQkRDoSY4zpFO0mAhE5taXl+05UEzPqKqkkldzEkC7KNsaYbi+Uo9mtze4nAxOARcAZYYmom/PUV+Illbg46zIxxkSHUJqGLmz+WEQGAn8OW0TdnMdXRXVcaqTDMMaYTnMgs6oUAsM7O5BDRUJDNXWetEiHYYwxnSaUPoK/4K4mBpc4RuOuMI5JSf4q6hMGRDoMY4zpNKH0ERQ0u98APK+q/w5TPN1ecsBLQ7zVCIwx0SOURPAyUKuqfgAR8YhIqqp6wxta95Sm1fgTbZwhY0z0COnKYiCl2eMU4L3whNPNNdSRiA9NtIvJjDHRI5REkNx8esrg/dg8babOfQxqs5MZY6JIKImgWkTGNj4QkXFATfhC6r7qqncDEGdzERhjokgofQS3AC+JyDbcVJX9cFNXxhxvRRlJgCfF+giMMdEjlAvKForIMGBocNFaVfWFN6zuqaZqN72A+FSrERhjokcok9f/EEhT1RWqugJIF5H/DH9o3U9dcHayxFSbi8AYEz1C6SO4NjhDGQCquhu4NnwhdV/1wUlpktKtRmCMiR6hJAJP80lpRMQDJIYvpO6rwSalMcZEoVA6i98GXhCRvwUfXw/MDV9I3Vfj7GSpGdY0ZIyJHqHUCH4JvA/cELwtZ+8LzFolIpNEZK2IrBOR21p4/k8isjR4+0pEylraTnehtZXUq4eMNBtiwhgTPUI5ayggIp8DhwOXA9nAK+29LtiE9ChwNm7E0oUiMltVVzXb9k+arX8zMKbDe9CV6iqoJJUeyTY7mTEmerSaCETkKGB68LYLeAFAVSeGuO0JwDpVXR/c3ixgKrCqlfWnA3eFuO2IiKurpJpUsjwHMnq3McZ0T20d0dbgZiG7QFVPUdW/AP4ObDsX2NLscWFw2X5EZDAwBNcE1dLz14lIgYgUFBcXdyCEzuXxVVJjk9IYY6JMW4ngYmA7MF9EHheRM3FXFofDNODlxhFO96WqM1V1vKqOz8nJCVMI7Yv3VVEbZ/0Dxpjo0moiUNXXVXUaMAyYjxtqoo+IPCYi54Sw7a3AwGaP84LLWjINeD60kCMn0W+zkxljok+7jd2qWq2qzwXnLs4DluDOJGrPQuBIERkiIom4g/3sfVcKDl/RC/i0Q5FHQJK/Gl+8DUFtjIkuHer1VNXdwWaaM0NYtwG4CXgHWA28qKorReReEZnSbNVpwCxV1Za2052kajUNCZYIjDHRJZQLyg6Yqs4B5uyz7M59Ht8dzhg6jSqpWkPAJqUxxkQZOw8yVA11JNAAiTYpjTEmulgiCJHPG7zo2SalMcZEGUsEIaquDM5OZpPSGGOijCWCENVUuhpBfIrVCIwx0cUSQYhqgzWChDRLBMaY6GKJIER11cFJaSwRGGOijCWCEPkaZydL6xXhSIwxpnNZIgiRTUpjjIlWlghCFAgmgrRMqxEYY6KLJYIQaV0ltZpgs5MZY6KOJYIQSV0lVaSSGG8fmTEmuthRLURxvkq8EtJUzcYYc0ixRBCieF8lNTYpjTEmClkiCFFCg01KY4yJTpYIQpTkr6beEoExJgpZIghRsr8aX7wNQW2MiT6WCEKUol78NimNMSYKWSIIhSpp6kUTbQhqY0z0sUQQAn9dNfESgCRrGjLGRB9LBCFonJRGki0RGGOijyWCEHiDicCTYgPOGWOijyWCENQ0TkqTan0ExpjoY4kgBHVVbuRRm53MGBONLBGEoN7r5itOSrchqI0x0ccSQQgagtNUpqRbH4ExJvpYIgiBv7YCsNnJjDHRyRJBCALBRJDRo3eEIzHGmM5niSAEUleBV5NISkyMdCjGGNPpLBGEQOorqZYURCTSoRhjTKezRBACT30VXrEhqI0x0ckSQQgSGqqojUuNdBjGGBMWYU0EIjJJRNaKyDoRua2VdS4XkVUislJEngtnPAfKzU5mQ1AbY6JTfLg2LCIe4FHgbKAQWCgis1V1VbN1jgRuB05W1d0i0idc8RyMZH8V1clZkQ7DGGPCIpw1ggnAOlVdr6r1wCxg6j7rXAs8qqq7AVR1ZxjjOWDJAS/+BOsjMMZEp3AmglxgS7PHhcFlzR0FHCUi/xaRz0RkUksbEpHrRKRARAqKi4vDFG7rUvESsElpjDFRKtKdxfHAkcDpwHTgcRHZ7/JdVZ2pquNVdXxOTk6XBqgBP+lag9qkNMaYKBXORLAVGNjscV5wWXOFwGxV9anqBuArXGLoNqqrKogTtUlpjDFRK5yJYCFwpIgMEZFEYBowe591XsfVBhCRbFxT0fowxtRhTbOT2RDUxpjoFLZEoKoNwE3AO8Bq4EVVXSki94rIlOBq7wAlIrIKmA/cqqol4YrpQNRUuEQQn2J9BMaY6BS200cBVHUOMGefZXc2u6/AT4O3bqm2ys1FkJBqNQJjTHSKdGdxt1db7RJBYppNSmOMiU6WCNrh8zZOSmM1AmNMdLJE0A6/181FkJJhNQJjTHSyRNCOQI2rEaRl2qQ0xpjoFNbO4qhQF5ymMt3OGjKms/l8PgoLC6mtrY10KFEjOTmZvLw8EhISQn6NJYL21FVSRQrpHvuojOlshYWFZGRkkJ+fbxM/dQJVpaSkhMLCQoYMGRLy66xpqB2e+kq82FwExoRDbW0tWVlZlgQ6iYiQlZXV4RqWJYJ2eHxV1NikNMaEjSWBznUgn6clgnYkNFRR57EhqI0x0csSQTuS/NXUWyIwJiqVlJQwevRoRo8eTb9+/cjNzd3zuL6+vs3XFhQU8KMf/aiLIg0v6wFtR3KgmrKEfpEOwxgTBllZWSxduhSAu+++m/T0dH7+85/veb6hoYH4+JYPk+PHj2f8+PFdEme4WSJoR4p62ZVgQ1AbE273vLGSVdsqOnWbIwZkcteFR3foNTNmzCA5OZklS5Zw8sknM23aNH784x9TW1tLSkoKTz75JEOHDuWDDz7gwQcf5M033+Tuu+9m8+bNrF+/ns2bN3PLLbccUrUFSwRtUFXS1IsmWiIwJpYUFhayYMECPB4PFRUVfPzxx8THx/Pee+/xq1/9ildeeWW/16xZs4b58+dTWVnJ0KFDufHGGzt0Ln8kWSJoQ22dj3SpBZudzJiw62jJPZwuu+wyPB4PAOXl5VxzzTV8/fXXiAg+n6/F10yePJmkpCSSkpLo06cPO3bsIC8vryvDPmDWWdyGqko38qik2IBzxsSStLSmE0R+/etfM3HiRFasWMEbb7zR6jn6SUlJe+57PB4aGhrCHmdnsUTQhurgpDSeZBtewphYVV5eTm5uLgBPPfVUZIMJE0sEbaipCs5OlmqJwJhY9Ytf/ILbb7+dMWPGHFKl/I4QN0nYoWP8+PFaUFDQJe+1dME7jP7X5aw9+2mGnvztLnlPY2LJ6tWrGT58eKTDiDotfa4iskhVWzzf1WoEbfAFZydLSu8Z4UiMMSZ8LBG0oWl2MksExpjoZYmgDf7gpDQpGTYpjTEmelkiaEOg1l3lmJZhNQJjTPSyRNCWukoCKniS7YIyY0z0skTQhri6CqolBWy8dGNMFLNE0AaPr4oasUlpjIlWEydO5J133tlr2Z///GduvPHGFtc//fTTaTx9/fzzz6esrGy/de6++24efPDBNt/39ddfZ9WqVXse33nnnbz33nsdDb/TWCJoQ3xDFTWe9EiHYYwJk+nTpzNr1qy9ls2aNYvp06e3+9o5c+bQs+eB9R/umwjuvfdezjrrrAPaVmewQefakNhQRb3HagTGdIm5t0HR8s7dZr9j4bzftvr0pZdeyh133EF9fT2JiYls3LiRbdu28fzzz/PTn/6UmpoaLr30Uu655579Xpufn09BQQHZ2dk88MADPP300/Tp04eBAwcybtw4AB5//HFmzpxJfX09RxxxBM888wxLly5l9uzZfPjhh9x///288sor3HfffVxwwQVceumlzJs3j5///Oc0NDRw3HHH8dhjj5GUlER+fj7XXHMNb7zxBj6fj5deeolhw4Z1ysdkNYI2JPm9+OKtRmBMtOrduzcTJkxg7ty5gKsNXH755TzwwAMUFBSwbNkyPvzwQ5YtW9bqNhYtWsSsWbNYunQpc+bMYeHChXueu/jii1m4cCFffvklw4cP54knnuCkk05iypQp/OEPf2Dp0qUcfvjhe9avra1lxowZvPDCCyxfvpyGhgYee+yxPc9nZ2ezePFibrzxxnabnzrCagRtSAlUUx4/KNJhGBMb2ii5h1Nj89DUqVOZNWsWTzzxBC+++CIzZ86koaGB7du3s2rVKkaOHNni6z/++GMuuugiUlNd68GUKVP2PLdixQruuOMOysrKqKqq4txzz20zlrVr1zJkyBCOOuooAK655hoeffRRbrnlFsAlFoBx48bx6quvHvS+N7IaQQtUlQZ/gFT1Eki0GoEx0Wzq1KnMmzePxYsX4/V66d27Nw8++CDz5s1j2bJlTJ48udWhp9szY8YMHnnkEZYvX85dd911wNtp1DjUdWcPcx0zNYInPtnAH/+1tsXnVMGvSiCg+FVpHIdvVVINYpPSGBPV0tPTmThxIt///veZPn06FRUVpKWl0aNHD3bs2MHcuXM5/fTTW339qaeeyowZM7j99ttpaGjgjTfe4PrrrwegsrKS/v374/P5ePbZZ/cMZ52RkUFlZeV+2xo6dCgbN25k3bp1e/oUTjvttLDsd3NhTQQiMgl4CPAAf1fV3+7z/AzgD8DW4KJHVPXv4YjltOq3mZL6tzaCBWm6CyKkeus4ctCAcIRjjOlGpk+fzkUXXcSsWbMYNmwYY8aMYdiwYQwcOJCTTz65zdeOHTuWK664glGjRtGnTx+OO+64Pc/dd999HH/88eTk5HD88cfvOfhPmzaNa6+9locffpiXX355z/rJyck8+eSTXHbZZXs6i2+44Ybw7HQzYRuGWkQ8wFfA2UAhsBCYrqqrmq0zAxivqjeFut0DHoZ6zVuw7IWOvSYuHk69FfrYMLnGhIMNQx0eHR2GOpw1ggnAOlVdHwxiFjAVWNXmq8Jl2GR3M8YYs5dwdhbnAluaPS4MLtvXJSKyTEReFpGBLW1IRK4TkQIRKSguLg5HrMYYE7MifdbQG0C+qo4E3gWebmklVZ2pquNVdXxOTk6XBmiMCa9DbZbE7u5APs9wJoKtQPMSfh5NncIAqGqJqtYFH/4dGBfGeIwx3UxycjIlJSWWDDqJqlJSUkJycnKHXhfOPoKFwJEiMgSXAKYB32m+goj0V9XtwYdTgNVhjMcY083k5eVRWFiINfl2nuTkZPLy8jr0mrAlAlVtEJGbgHdwp4/+Q1VXisi9QIGqzgZ+JCJTgAagFJgRrniMMd1PQkICQ4YMiXQYMS9sp4+GywGfPmqMMTGsrdNHI91ZbIwxJsIsERhjTIw75JqGRKQY2HSAL88GdnViOIeKWN1viN19t/2OLaHs92BVbfH8+0MuERwMESlorY0smsXqfkPs7rvtd2w52P22piFjjIlxlgiMMSbGxVoimBnpACIkVvcbYnffbb9jy0Htd0z1ERhjjNlfrNUIjDHG7MMSgTHGxLiYSQQiMklE1orIOhG5LdLxhIuI/ENEdorIimbLeovIuyLydfBvr0jGGA4iMlBE5ovIKhFZKSI/Di6P6n0XkWQR+UJEvgzu9z3B5UNE5PPg9/0FEUmMdKzhICIeEVkiIm8GH0f9fovIRhFZLiJLRaQguOygvucxkQiC02Y+CpwHjACmi8iIyEYVNk8Bk/ZZdhswT1WPBOYFH0ebBuBnqjoCOAH4YfB/HO37XgecoaqjgNHAJBE5Afgd8CdVPQLYDfwggjGG04/Ze9TiWNnviao6utm1Awf1PY+JRECzaTNVtR5onDYz6qjqR7iRXJubStOkP08D3+7SoLqAqm5X1cXB+5W4g0MuUb7v6lQFHyYEbwqcATTOih51+w0gInnAZNxcJoiIEAP73YqD+p7HSiIIddrMaNW32bwPRUDfSAYTbiKSD4wBPicG9j3YPLIU2Imb6e8boExVG4KrROv3/c/AL4BA8HEWsbHfCvxLRBaJyHXBZQf1PQ/nxDSmG1JVFZGoPWdYRNKBV4BbVLXCFRKdaN13VfUDo0WkJ/AaMCzCIYWdiFwA7FTVRSJyeqTj6WKnqOpWEekDvCsia5o/eSDf81ipEbQ7bWaU2yEi/cHNCocrOUYdEUnAJYFnVfXV4OKY2HcAVS0D5gMnAj1FpLGgF43f95OBKSKyEdfUewbwENG/36jq1uDfnbjEP4GD/J7HSiLYM21m8CyCacDsCMfUlWYD1wTvXwP8XwRjCYtg+/ATwGpV/e9mT0X1votITrAmgIikAGfj+kfmA5cGV4u6/VbV21U1T1Xzcb/n91X1SqJ8v0UkTUQyGu8D5wArOMjvecxcWSwi5+PaFBunzXwgwiGFhYg8D5yOG5Z2B3AX8DrwIjAIN4T35aq6b4fyIU1ETgE+BpbT1Gb8K1w/QdTuu4iMxHUOenAFuxdV9V4ROQxXUu4NLAGuUtW6yEUaPsGmoZ+r6gXRvt/B/Xst+DAeeE5VHxCRLA7iex4zicAYY0zLYqVpyBhjTCssERhjTIyzRGCMMTHOEoExxsQ4SwTGGBPjLBEYsw8R8QdHdmy8ddpAdSKS33xkWGO6Axtiwpj91ajq6EgHYUxXsRqBMSEKjgP/++BY8F+IyBHB5fki8r6ILBOReSIyKLi8r4i8Fpwr4EsROSm4KY+IPB6cP+BfwSuCjYkYSwTG7C9ln6ahK5o9V66qxwKP4K5UB/gL8LSqjgSeBR4OLn8Y+DA4V8BYYGVw+ZHAo6p6NFAGXBLm/TGmTXZlsTH7EJEqVU1vYflG3CQw64MD3BWpapaI7AL6q6ovuHy7qmaLSDGQ13yIg+AQ2e8GJxBBRH4JJKjq/eHfM2NaZjUCYzpGW7nfEc3HvvFjfXUmwiwRGNMxVzT7+2nw/gLcCJgAV+IGvwM3ZeCNsGfymB5dFaQxHWElEWP2lxKc8avR26raeAppLxFZhivVTw8uuxl4UkRuBYqB7wWX/xiYKSI/wJX8bwS2Y0w3Y30ExoQo2EcwXlV3RToWYzqTNQ0ZY0yMsxqBMcbEOKsRGGNMjLNEYIwxMc4SgTHGxDhLBMYYE+MsERhjTIz7/99Ck3TAPIDhAAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "glove = torchtext.vocab.GloVe(name=\"6B\", dim=100)\n",
        "\n",
        "model = Transformer(100).cuda()\n",
        "train_network(model, train_data, valid_data, batch_size = 512, num_epochs=50, learning_rate=2e-4)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "1aX290FTJNpp",
        "outputId": "de06145c-6797-4686-a290-3a89e6597af2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|█████████▉| 399999/400000 [00:15<00:00, 25065.24it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1; Loss 0.606626; Train Acc 0.733048; Val Acc 0.730913\n",
            "Epoch 2; Loss 0.464674; Train Acc 0.799429; Val Acc 0.790710\n",
            "Epoch 3; Loss 0.361335; Train Acc 0.845111; Val Acc 0.849973\n",
            "Epoch 4; Loss 0.329470; Train Acc 0.870985; Val Acc 0.859584\n",
            "Epoch 5; Loss 0.281582; Train Acc 0.882049; Val Acc 0.862787\n",
            "Epoch 6; Loss 0.274132; Train Acc 0.895075; Val Acc 0.880940\n",
            "Epoch 7; Loss 0.269596; Train Acc 0.907209; Val Acc 0.882541\n",
            "Epoch 8; Loss 0.243882; Train Acc 0.915418; Val Acc 0.891084\n",
            "Epoch 9; Loss 0.188199; Train Acc 0.924518; Val Acc 0.898025\n",
            "Epoch 10; Loss 0.180107; Train Acc 0.924518; Val Acc 0.892152\n",
            "Epoch 11; Loss 0.194874; Train Acc 0.945396; Val Acc 0.905499\n",
            "Epoch 12; Loss 0.193992; Train Acc 0.954675; Val Acc 0.908703\n",
            "Epoch 13; Loss 0.122393; Train Acc 0.960207; Val Acc 0.909770\n",
            "Epoch 14; Loss 0.125910; Train Acc 0.963776; Val Acc 0.914042\n",
            "Epoch 15; Loss 0.086164; Train Acc 0.974304; Val Acc 0.912974\n",
            "Epoch 16; Loss 0.074600; Train Acc 0.977516; Val Acc 0.910838\n",
            "Epoch 17; Loss 0.064235; Train Acc 0.983048; Val Acc 0.911906\n",
            "Epoch 18; Loss 0.063073; Train Acc 0.986795; Val Acc 0.911372\n",
            "Epoch 19; Loss 0.052746; Train Acc 0.989293; Val Acc 0.919381\n",
            "Epoch 20; Loss 0.054613; Train Acc 0.992862; Val Acc 0.913508\n",
            "Epoch 21; Loss 0.034501; Train Acc 0.993398; Val Acc 0.914576\n",
            "Epoch 22; Loss 0.016761; Train Acc 0.962348; Val Acc 0.890550\n",
            "Epoch 23; Loss 0.046506; Train Acc 0.986974; Val Acc 0.906033\n",
            "Epoch 24; Loss 0.017606; Train Acc 0.992148; Val Acc 0.908169\n",
            "Epoch 25; Loss 0.013306; Train Acc 0.996074; Val Acc 0.908703\n",
            "Epoch 26; Loss 0.014180; Train Acc 0.997680; Val Acc 0.909770\n",
            "Epoch 27; Loss 0.007950; Train Acc 0.998037; Val Acc 0.908703\n",
            "Epoch 28; Loss 0.006097; Train Acc 0.997680; Val Acc 0.903364\n",
            "Epoch 29; Loss 0.022907; Train Acc 0.999465; Val Acc 0.910304\n",
            "Epoch 30; Loss 0.025741; Train Acc 0.997680; Val Acc 0.909237\n",
            "Epoch 31; Loss 0.006182; Train Acc 0.999108; Val Acc 0.912440\n",
            "Epoch 32; Loss 0.003343; Train Acc 0.999286; Val Acc 0.910304\n",
            "Epoch 33; Loss 0.003397; Train Acc 0.999465; Val Acc 0.910304\n",
            "Epoch 34; Loss 0.007410; Train Acc 0.999108; Val Acc 0.911906\n",
            "Epoch 35; Loss 0.004706; Train Acc 0.999108; Val Acc 0.906567\n",
            "Epoch 36; Loss 0.004008; Train Acc 0.999286; Val Acc 0.906567\n",
            "Epoch 37; Loss 0.001708; Train Acc 0.999108; Val Acc 0.901762\n",
            "Epoch 38; Loss 0.004174; Train Acc 0.998751; Val Acc 0.906033\n",
            "Epoch 39; Loss 0.002093; Train Acc 0.998751; Val Acc 0.907635\n",
            "Epoch 40; Loss 0.024539; Train Acc 0.999643; Val Acc 0.906033\n",
            "Epoch 41; Loss 0.004488; Train Acc 0.998751; Val Acc 0.913508\n",
            "Epoch 42; Loss 0.007173; Train Acc 0.999465; Val Acc 0.909237\n",
            "Epoch 43; Loss 0.001927; Train Acc 0.999108; Val Acc 0.912440\n",
            "Epoch 44; Loss 0.003032; Train Acc 0.998751; Val Acc 0.910304\n",
            "Epoch 45; Loss 0.003709; Train Acc 0.996966; Val Acc 0.907635\n",
            "Epoch 46; Loss 0.023154; Train Acc 0.993576; Val Acc 0.899626\n",
            "Epoch 47; Loss 0.048645; Train Acc 0.997502; Val Acc 0.903897\n",
            "Epoch 48; Loss 0.003300; Train Acc 0.998394; Val Acc 0.912974\n",
            "Epoch 49; Loss 0.022490; Train Acc 0.998037; Val Acc 0.903364\n",
            "Epoch 50; Loss 0.009092; Train Acc 0.999643; Val Acc 0.910304\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "glove = torchtext.vocab.GloVe(name=\"6B\", dim=50)\n",
        "\n",
        "model = Transformer(50).cuda()\n",
        "train_network(model, train_data, valid_data, batch_size = 350, num_epochs=50, learning_rate=2e-4)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "aDXHwY2RJvKs",
        "outputId": "7c2c1491-297b-4369-ff73-d5cff37b4dd7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1; Loss 0.654704; Train Acc 0.689329; Val Acc 0.693006\n",
            "Epoch 2; Loss 0.722337; Train Acc 0.772484; Val Acc 0.768820\n",
            "Epoch 3; Loss 0.180373; Train Acc 0.812455; Val Acc 0.816337\n",
            "Epoch 4; Loss 0.177508; Train Acc 0.833333; Val Acc 0.832355\n",
            "Epoch 5; Loss 0.140599; Train Acc 0.841006; Val Acc 0.830219\n",
            "Epoch 6; Loss 0.921190; Train Acc 0.795682; Val Acc 0.785905\n",
            "Epoch 7; Loss 0.189782; Train Acc 0.849036; Val Acc 0.853711\n",
            "Epoch 8; Loss 0.385022; Train Acc 0.867773; Val Acc 0.865456\n",
            "Epoch 9; Loss 0.272811; Train Acc 0.879907; Val Acc 0.868660\n",
            "Epoch 10; Loss 0.346809; Train Acc 0.885439; Val Acc 0.869728\n",
            "Epoch 11; Loss 0.500374; Train Acc 0.873305; Val Acc 0.862253\n",
            "Epoch 12; Loss 0.156995; Train Acc 0.862598; Val Acc 0.849439\n",
            "Epoch 13; Loss 0.191498; Train Acc 0.887045; Val Acc 0.868660\n",
            "Epoch 14; Loss 0.460623; Train Acc 0.894540; Val Acc 0.876668\n",
            "Epoch 15; Loss 0.119889; Train Acc 0.897038; Val Acc 0.876135\n",
            "Epoch 16; Loss 0.423029; Train Acc 0.896859; Val Acc 0.882541\n",
            "Epoch 17; Loss 0.174935; Train Acc 0.887580; Val Acc 0.872931\n",
            "Epoch 18; Loss 0.186985; Train Acc 0.906138; Val Acc 0.883609\n",
            "Epoch 19; Loss 0.043564; Train Acc 0.908101; Val Acc 0.884677\n",
            "Epoch 20; Loss 0.092825; Train Acc 0.910778; Val Acc 0.885745\n",
            "Epoch 21; Loss 0.277064; Train Acc 0.913633; Val Acc 0.888948\n",
            "Epoch 22; Loss 0.221835; Train Acc 0.916131; Val Acc 0.889482\n",
            "Epoch 23; Loss 0.030231; Train Acc 0.918808; Val Acc 0.894287\n",
            "Epoch 24; Loss 0.029893; Train Acc 0.922020; Val Acc 0.898025\n",
            "Epoch 25; Loss 0.105340; Train Acc 0.922020; Val Acc 0.892152\n",
            "Epoch 26; Loss 0.035608; Train Acc 0.929336; Val Acc 0.887880\n",
            "Epoch 27; Loss 0.135327; Train Acc 0.936652; Val Acc 0.898025\n",
            "Epoch 28; Loss 0.287256; Train Acc 0.935582; Val Acc 0.895889\n",
            "Epoch 29; Loss 0.224987; Train Acc 0.931299; Val Acc 0.893753\n",
            "Epoch 30; Loss 0.045671; Train Acc 0.937009; Val Acc 0.894821\n",
            "Epoch 31; Loss 0.037950; Train Acc 0.942363; Val Acc 0.893219\n",
            "Epoch 32; Loss 0.288393; Train Acc 0.938437; Val Acc 0.893753\n",
            "Epoch 33; Loss 0.017265; Train Acc 0.946110; Val Acc 0.895355\n",
            "Epoch 34; Loss 0.021517; Train Acc 0.950571; Val Acc 0.899626\n",
            "Epoch 35; Loss 0.025768; Train Acc 0.955211; Val Acc 0.900160\n",
            "Epoch 36; Loss 0.052420; Train Acc 0.954675; Val Acc 0.896957\n",
            "Epoch 37; Loss 0.013086; Train Acc 0.965739; Val Acc 0.903364\n",
            "Epoch 38; Loss 0.006588; Train Acc 0.970735; Val Acc 0.903897\n",
            "Epoch 39; Loss 0.004795; Train Acc 0.975018; Val Acc 0.903897\n",
            "Epoch 40; Loss 0.038165; Train Acc 0.978051; Val Acc 0.907635\n",
            "Epoch 41; Loss 0.005259; Train Acc 0.972341; Val Acc 0.897491\n",
            "Epoch 42; Loss 0.065717; Train Acc 0.973233; Val Acc 0.898025\n",
            "Epoch 43; Loss 0.247567; Train Acc 0.895610; Val Acc 0.854778\n",
            "Epoch 44; Loss 0.016278; Train Acc 0.925589; Val Acc 0.879338\n",
            "Epoch 45; Loss 0.010562; Train Acc 0.957530; Val Acc 0.898025\n",
            "Epoch 46; Loss 0.002111; Train Acc 0.972698; Val Acc 0.909770\n",
            "Epoch 47; Loss 0.027657; Train Acc 0.975196; Val Acc 0.900694\n",
            "Epoch 48; Loss 0.021859; Train Acc 0.981442; Val Acc 0.911372\n",
            "Epoch 49; Loss 0.004001; Train Acc 0.983940; Val Acc 0.904431\n",
            "Epoch 50; Loss 0.021456; Train Acc 0.984475; Val Acc 0.899092\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "glove = torchtext.vocab.GloVe(name=\"6B\", dim=50)\n",
        "\n",
        "model = Transformer(50).cuda()\n",
        "train_network(model, train_data, valid_data, batch_size = 512, num_epochs=50, learning_rate=2e-5)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "7gbhVEKLK57C",
        "outputId": "dfa29dbd-75e7-4152-886c-7b0e80092092"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1; Loss 0.689933; Train Acc 0.548537; Val Acc 0.543513\n",
            "Epoch 2; Loss 0.672504; Train Acc 0.625803; Val Acc 0.614522\n",
            "Epoch 3; Loss 0.647038; Train Acc 0.680764; Val Acc 0.661506\n",
            "Epoch 4; Loss 0.631611; Train Acc 0.713241; Val Acc 0.694074\n",
            "Epoch 5; Loss 0.618223; Train Acc 0.740186; Val Acc 0.719167\n",
            "Epoch 6; Loss 0.582881; Train Acc 0.756959; Val Acc 0.737854\n",
            "Epoch 7; Loss 0.570189; Train Acc 0.767844; Val Acc 0.754405\n",
            "Epoch 8; Loss 0.553642; Train Acc 0.774268; Val Acc 0.753337\n",
            "Epoch 9; Loss 0.528205; Train Acc 0.782834; Val Acc 0.768286\n",
            "Epoch 10; Loss 0.500484; Train Acc 0.790328; Val Acc 0.776829\n",
            "Epoch 11; Loss 0.458205; Train Acc 0.796217; Val Acc 0.783769\n",
            "Epoch 12; Loss 0.467597; Train Acc 0.801035; Val Acc 0.793380\n",
            "Epoch 13; Loss 0.447716; Train Acc 0.807637; Val Acc 0.804592\n",
            "Epoch 14; Loss 0.463254; Train Acc 0.812991; Val Acc 0.809397\n",
            "Epoch 15; Loss 0.378181; Train Acc 0.818701; Val Acc 0.823812\n",
            "Epoch 16; Loss 0.437358; Train Acc 0.823876; Val Acc 0.827549\n",
            "Epoch 17; Loss 0.413586; Train Acc 0.828694; Val Acc 0.828617\n",
            "Epoch 18; Loss 0.372921; Train Acc 0.827445; Val Acc 0.833956\n",
            "Epoch 19; Loss 0.409369; Train Acc 0.832620; Val Acc 0.838761\n",
            "Epoch 20; Loss 0.383202; Train Acc 0.837794; Val Acc 0.838761\n",
            "Epoch 21; Loss 0.353840; Train Acc 0.841363; Val Acc 0.844100\n",
            "Epoch 22; Loss 0.356255; Train Acc 0.843683; Val Acc 0.844634\n",
            "Epoch 23; Loss 0.381635; Train Acc 0.845824; Val Acc 0.846236\n",
            "Epoch 24; Loss 0.405021; Train Acc 0.847787; Val Acc 0.851575\n",
            "Epoch 25; Loss 0.369322; Train Acc 0.850107; Val Acc 0.851575\n",
            "Epoch 26; Loss 0.326135; Train Acc 0.851535; Val Acc 0.852643\n",
            "Epoch 27; Loss 0.372033; Train Acc 0.853676; Val Acc 0.854778\n",
            "Epoch 28; Loss 0.361853; Train Acc 0.853319; Val Acc 0.855312\n",
            "Epoch 29; Loss 0.307739; Train Acc 0.858137; Val Acc 0.857982\n",
            "Epoch 30; Loss 0.291228; Train Acc 0.858851; Val Acc 0.859050\n",
            "Epoch 31; Loss 0.376217; Train Acc 0.857066; Val Acc 0.855846\n",
            "Epoch 32; Loss 0.358301; Train Acc 0.859208; Val Acc 0.858516\n",
            "Epoch 33; Loss 0.352356; Train Acc 0.864026; Val Acc 0.859050\n",
            "Epoch 34; Loss 0.385025; Train Acc 0.865096; Val Acc 0.863855\n",
            "Epoch 35; Loss 0.343388; Train Acc 0.862598; Val Acc 0.862253\n",
            "Epoch 36; Loss 0.267886; Train Acc 0.864026; Val Acc 0.862787\n",
            "Epoch 37; Loss 0.330816; Train Acc 0.866702; Val Acc 0.863321\n",
            "Epoch 38; Loss 0.338146; Train Acc 0.868844; Val Acc 0.868126\n",
            "Epoch 39; Loss 0.358611; Train Acc 0.870093; Val Acc 0.867592\n",
            "Epoch 40; Loss 0.290511; Train Acc 0.869022; Val Acc 0.867592\n",
            "Epoch 41; Loss 0.338844; Train Acc 0.869557; Val Acc 0.867058\n",
            "Epoch 42; Loss 0.313359; Train Acc 0.871699; Val Acc 0.869194\n",
            "Epoch 43; Loss 0.323364; Train Acc 0.872413; Val Acc 0.870262\n",
            "Epoch 44; Loss 0.314535; Train Acc 0.873840; Val Acc 0.876135\n",
            "Epoch 45; Loss 0.289654; Train Acc 0.875268; Val Acc 0.873465\n",
            "Epoch 46; Loss 0.266824; Train Acc 0.875803; Val Acc 0.872397\n",
            "Epoch 47; Loss 0.313121; Train Acc 0.876517; Val Acc 0.877202\n",
            "Epoch 48; Loss 0.285004; Train Acc 0.877766; Val Acc 0.873465\n",
            "Epoch 49; Loss 0.300390; Train Acc 0.877766; Val Acc 0.877736\n",
            "Epoch 50; Loss 0.330813; Train Acc 0.879015; Val Acc 0.877202\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3dd3zV9fX48dfJ3iE7IYMMSMKegoooWBVQiwO1om3Vtlqtq61a7fj122H7rZ1+tdZdtcW6R20d4AAZyl6yAgmQBSF77+T9++PexBBukpvk3twk9zwfDx7c+8nnc+/5YLznvtd5izEGpZRS7svD1QEopZRyLU0ESinl5jQRKKWUm9NEoJRSbk4TgVJKuTlNBEop5eY0ESi3ISLvi8gNjj5XqZFOdB2BGs5EpLbL0wCgCWizPv+uMebFoY9qcEQkBPgVcCUQDpwE/gM8aIwpdWVsyj1pi0ANa8aYoI4/QB7w1S7HOpOAiHi5Lkr7iYgP8DEwGVgChABnAWXA3AG83oi4bzW8aSJQI5KILBSRAhG5X0SKgOdEJExE/isiJSJSYX2c0OWatSLyHevjG0Vkg4j80XruURFZOsBzU0RknYjUiMhHIvKYiKzsIfRvAknAFcaY/caYdmNMsTHm18aY96yvZ0RkfJfXf15EHuzlvg+IyKVdzvey/hvMsj4/U0Q+E5FKEdktIgsH+++vRhdNBGoki8XStTIOuAXL7/Nz1udJQAPw116unwdkAZHA74FnRUQGcO6/gC1ABPAL4Bu9vOcFwAfGmNpezulL9/t+CVjR5eeLgVJjzA4RiQfeBR60XnMv8IaIRA3i/dUoo4lAjWTtwP8YY5qMMQ3GmDJjzBvGmHpjTA3wG+C8Xq7PNcY8bYxpA14A4oCY/pwrIknAGcDPjTHNxpgNwDu9vGcEcKJ/t3maU+4bSyJaJiIB1p9fhyU5AHwdeM8Y85619fEhsA24eJAxqFFEE4EayUqMMY0dT0QkQESeFJFcEakG1gFjRMSzh+uLOh4YY+qtD4P6ee5YoLzLMYD8XmIuw5JEBuOU+zbGZAMHgK9ak8EyLMkBLK2Gq63dQpUiUgmc44AY1CiiA01qJOs+5e0eIAOYZ4wpEpEZwE6gp+4eRzgBhItIQJdkkNjL+R8BD4pIoDGmrodz6rHMkOoQCxR0eW5rql9H95AHsN+aHMCSlP5pjLm5j/tQbkxbBGo0CcYyLlApIuHA/zj7DY0xuVi6Wn4hIj4ichbw1V4u+SeWD+c3RCRTRDxEJEJEfiIiHd01u4DrRMRTRJbQe/dWh5eBi4Db+LI1ALASS0thsfX1/KwDzgk2X0W5JU0EajR5GPAHSoFNwAdD9L7X8+UU0AeBV7CsdziNMaYJy4DxQeBDoBrLQHMksNl62t1Ykkml9bXf7isAY8wJ4HPgbOv7dxzPBy4DfgKUYElC96H/76sudEGZUg4mIq8AB40xTm+RKOUI+q1AqUESkTNEJM3azbMEyzfwPr/FKzVc6GCxUoMXC7yJZWpoAXCbMWana0NSyn7aNaSUUm5Ou4aUUsrNjbiuocjISJOcnOzqMJRSakTZvn17qTHGZmmREZcIkpOT2bZtm6vDUEqpEUVEcnv6mVO7hkRkiYhkiUi2iDxg4+d/EZFd1j+HrMvflVJKDSGntQis9V0eAy7EMpNiq4i8Y4zZ33GOMeYHXc6/E5jprHiUUkrZ5swWwVwg2xhzxBjTjGUJ/GW9nL+CLysmKqWUGiLOTATxnFqFscB67DQiMg5IAT5xYjxKKaVsGC7TR68FXrfWej+NiNwiIttEZFtJSckQh6aUUqObMxNBIaeW402wHrPlWnrpFjLGPGWMmWOMmRMVpRsrKaWUIzkzEWwFJlj3c/XB8mF/2s5NIpIJhGGpnKiUUmqIOS0RGGNagTuAVVh2T3rVGLNPRH4lIsu6nHot8LJxcq2LnJJaHvrgIFpSQymlTuXUBWXGmPeA97od+3m3579wZgwd1hws5vG1OUQH+3LT/JSheEullBoRhstgsdN9+5wULpgYw2/ePcCOvApXh6OUUsOG2yQCEeFPV08nbowfd7y4g4q6ZleHpJRSw4LbJAKA0ABv/nbdbEprm/n+K7tob9fxAqWUcqtEADA1IZSff3USnx4q4bE12a4ORymlXM7tEgHA9fOSuHzGWP7y0SE2Zpe6OhyllHIpt0wEIsJvrphKalQQd7+8k5PVja4OSSmlXMYtEwFAoK8Xj18/i7qmNu78105a29pdHZJSSrmE2yYCgAkxwfz2yilsOVbOS1vz+75AKaVGIbdOBACXz4jnjOQwHv34MI0tNmveKaXUqOb2iUBEuG9xJsU1Tfzj82OuDkcppYac2ycCgLkp4ZybHsXja3OoaWxxdThKKTWkNBFY3XdRBhX1LTy74airQ1FKqSGlicBqakIoS6fE8sz6o1p+QinlVjQRdPHDC9Opa27liU9zXB2KUkoNGU0EXUyICeaKmfE8/9kxXWSmlHIbmgi6+f5X0mlrNzz6yWFXh6KUUkNCE0E3SREBXDs3kZe35JNXVu/qcJRSyuk0Edhw5/kT8PQQHv74kKtDUUopp9NEYENMiB83nJ3MWzsLOXSyxtXhKKWUU2ki6MGt56UR4ufN7S/uoKpBF5kppUYvTQQ9CA/04fGvz+JYWR3fe3E7za1anVQpNTppIujF2WmR/O+V09iYXcZP3/oCY3RrS6XU6OPl6gCGu6tmJ5BXXs8jHx9mXEQAd5w/wdUhKaWUQ2kisMMPLphAXlkdf1x9iMTwAC6bEe/qkJRSymE0EdhBRHjoqmkcr2zkvtf2MHaMP2ckh7s6LKWUcggdI7CTr5cnT35jNvFh/tz8j20cLa1zdUhKKeUQmgj6ISzQh+duPAMPEb7+zGbyy3XlsVJq5NNE0E/JkYG8cNNcahpbWPH0JgoqNBkopUY2TQQDMDUhlJXfmUdVgyUZFFY2uDokpZQaME0EAzQtYQwrvz2PyroWVjy1ieOaDJRSI5QmgkGYnjiGf3x7LhV1zax4ehMnqjQZKKVGHk0EgzQzKYwXvj2XstpmVjy1iaIq3dBGKTWyaCJwgFlJYbzwrbmU1jbzree3aikKpdSIoonAQWaPC+Nnl0xk/4lqduRVujocpZSymyYCB7p0+lj8vT15fXuBq0NRSim7aSJwoCBfL5ZOjeW/u4/T0Nzm6nCUUsoumggc7KrZCdQ0tbJ6f5GrQ1FKKbtoInCwM1MiSAjz1+4hpdSI4dREICJLRCRLRLJF5IEezrlGRPaLyD4R+Zcz4xkKHh7C8lkJbMgu1UVmSqkRwWmJQEQ8gceApcAkYIWITOp2zgTgx8B8Y8xk4PvOimcoLZ+VgDHw5g5tFSilhj9ntgjmAtnGmCPGmGbgZeCybufcDDxmjKkAMMYUOzGeIZMUEcC8lHBe316gawqUUsOeMxNBPJDf5XmB9VhX6UC6iGwUkU0issTWC4nILSKyTUS2lZSUOClcx7p6TiLHyurZnlvh6lCUUqpXrh4s9gImAAuBFcDTIjKm+0nGmKeMMXOMMXOioqKGOMSBWTollgAfT17bpt1DSqnhzZmJoBBI7PI8wXqsqwLgHWNMizHmKHAIS2IY8QJ9vbh4ahzvfnGC+uZWV4ejlFI9cmYi2ApMEJEUEfEBrgXe6XbO21haA4hIJJauoiNOjGlIXT07gdqmVlbt0zUFSqnhy2mJwBjTCtwBrAIOAK8aY/aJyK9EZJn1tFVAmYjsB9YA9xljypwV01A7IzmcpPAA7R5SSg1rXs58cWPMe8B73Y79vMtjA/zQ+mfU6VhT8JePDlFQUU9CWICrQ1JKqdO4erB41Fs+2zJR6s0d3YdHlFJqeNBE4GQJYQGcnRbB69sLaG1rd3U4Sil1Gk0EQ+AbZ44jr7yeO1/aSXOrJgOl1PCiiWAILJ0ax88umcj7e4v47j+30diiJaqVUsOHJoIh8p0Fqfz2iqmsPVTCTc9tpa5J1xYopYYHTQRD6Lp5Sfz5mulsOVbON57dTFVDi6tDUkopTQRD7YqZCTx23Uy+KKziuqc3UV7X7OqQlFJuThOBCyyZEsdT35xDdnEtX3vyc0prm1wdklLKjWkicJFFGdE8f9Nc8ivq+d7KHbTo1FKllItoInChs9IieGj5NLYcK+fX/93v6nCUUm7KqSUmVN8umxHPvuPVPLXuCFPGhnLNGYl9X6SUUg6kLYJh4EeLMzhnfCQ/e3svO/N0Ixul1NDSRDAMeHl68OiKmcSE+nLryu0UVze6OiSllBvRRDBMhAX68NQ35lDd0MptL+7QUhRKqSGjiWAYmRgXwh+unsb23Ap+8Z99rg5HKeUmNBEMM5dOG8ttC9P41+Y83tiuG9oopZxPE8EwdO9FGcxMGsMfVmVpgTqllNNpIhiGPD2E+5dkUlTdyMpNua4ORyk1ymkiGKbOTI1gwYRIHluTTU2jFqdTSjmPJoJh7N6LMqiob+HvG465OhSl1CimiWAYm544hsWTY3hm/REqtEqpUspJNBEMc/dclEFtcytPrMtxdShKqVFKE8Ewlx4TzBUz4nnhs2Oc1BXHSikn0EQwAnz/gnRa2wx//STb1aEopUYhTQQjQFJEANfOTeSlLXnkldW7Ohyl1CijiWCEuPP8CXh6CA9/fMjVoSilRhlNBCNETIgfN5ydzNs7Czl8ssbV4SilRhFNBCPIreelEeDjxV8+0laBUspxNBGMIOGBPtw0P5n39xaRXaytAqWUY2giGGFump+Cn5cnf1uj6wqUUo6hiWCECQ/04bp5Sfx793Hyy3UGkVJq8DQRjEC3nJuKpwhPfKqtAqXU4GkiGIFiQvy4ak4Cr20r0NXGSqlB00QwQt16bhptxvDM+iOuDkUpNcJpIhihkiICWDZ9LC9uztPKpEqpQdFEMIJ9b2Ea9c1tPLfxqKtDUUqNYJoIRrAJMcEsmRzL858d013MlFIDpolghLt90XiqG1tZuSnP1aEopUYopyYCEVkiIlkiki0iD9j4+Y0iUiIiu6x/vuPMeEajqQmhnJsexbMbjtDY0ubqcJRSI5DTEoGIeAKPAUuBScAKEZlk49RXjDEzrH+ecVY8o9kdi8ZTWtvMy1u0VaCU6j9ntgjmAtnGmCPGmGbgZeAyJ76f25qbEs4ZyWE8vf4o7e3G1eEopUYYZyaCeCC/y/MC67HulovIHhF5XUQSbb2QiNwiIttEZFtJSYkzYh3xrpuXRGFlA7sKKl0dilJqhHH1YPF/gGRjzDTgQ+AFWycZY54yxswxxsyJiooa0gBHivMzYvDyEFbtK3J1KEqpEcauRCAigSLiYX2cLiLLRMS7j8sKga7f8BOsxzoZY8qMMU3Wp88As+0LW3UXGuDNmakRrN53EmO0e0gpZT97WwTrAD8RiQdWA98Anu/jmq3ABBFJEREf4Frgna4niEhcl6fLgAN2xqNsWDw5hqOldeSU1Lo6FKXUCGJvIhBjTD1wJfA3Y8zVwOTeLjDGtAJ3AKuwfMC/aozZJyK/EpFl1tPuEpF9IrIbuAu4cSA3oSwunBQLwKp9J10ciVJqJPGy8zwRkbOA64FvW4959nWRMeY94L1ux37e5fGPgR/bGYPqQ2yoH9MTx7B6XxG3Lxrv6nCUUiOEvS2C72P5wH7L+q0+FVjjvLDUQF00KYbdBVWcqGpwdShKqRHCrkRgjPnUGLPMGPOQddC41Bhzl5NjUwOweLKle+jD/do9pJSyj72zhv4lIiEiEgjsBfaLyH3ODU0NxPjoIFKjAnUaqVLKbvZ2DU0yxlQDlwPvAylYZg6pYWjx5Fg2HSmnql4rkiql+mZvIvC2rhu4HHjHGNMC6GT1YeqiSTG0tRs+PqjdQ0qpvtmbCJ4EjgGBwDoRGQdUOysoNTjTE8YQE+LLap1GqpSyg72DxY8YY+KNMRcbi1xgkZNjUwPk4SFcOCmGTw+VaGlqpVSf7B0sDhWRP3cUfhORP2FpHahhavHkWBpa2lh/uNTVoSilhjl7u4b+DtQA11j/VAPPOSsoNXjzUiII9vPqcfZQcXUjv3v/IPnl9UMcmVJquLF3ZXGaMWZ5l+e/FJFdzghIOYaPlwdfyYzm4wMnaW1rx8vzy5y/el8RD7z5BeV1zWw9Vs6r3z0LTw9xYbRKKVeyt0XQICLndDwRkfmALl0d5i6aHEtFfQtbj1UAUN/cyo/f3MMt/9zO2DF+3HNhOttzK3hu41G7Xu/D/Sd5/4sTzgxZKeUC9rYIbgX+ISKh1ucVwA3OCUk5ynnpUfh4ebB6fxEBPp58/5VdHCur49bz0vjhhel4ewq7C6r4w6osFmVGkxYV1ONrrT9cwq0rt5MQ5s/SqXE9nqeUGnnsnTW02xgzHZgGTDPGzATOd2pkatACfb1YMD6S17YVsPzxz2hqaeOlm8/kgaWZ+Hh5ICL89oop+Hl7ct9ru2nrYZvLQydr+N7KHbQbQ355PU2tOhNJqdGkXzuUGWOqrSuMAX7ohHiUg106PY7aplYunhrH+98/lzNTI075eXSIH7+6bDI78ip5dsOR064vqWnipue24ufjyY8WZ9JuILdMB5iVGk3s7RqyRUcXR4DLZ8QzIzGMlMieZ/sumz6Wd/ec4I+rD3F+ZjTjo4MBaGxp4+Z/bKOsrolXv3sWHiI89AHkFNeSHhM8VLeglHKywexZrCUmRgAR6TUJdJzz4BVTCPDx5N7X9tDWbmhvN/zw1V3sLqjk/66dybSEMaRGWV5Hd0BTanTptUUgIjXY/sAXwN8pESmXiA7245fLJnP3y7t4ev0RqhpaeO+LIn568cTO0tYBPl7Ej/Enp6TOxdEqpRyp10RgjNH2vxtZNn0s731xgj+syqKt3XDdvCS+syDllHNSowK1RaDUKDOYriE1yogID14+lbAAH85Lj+KXyyYjcupQUFpUEDnFtRijPYNKjRaDGSxWo1BUsC/rfrQQPy9PPGysNk6LDqKuuY2T1U3Ehvq5IEKllKNpi0CdJsDHy2YSAEjTAWOlRh1NBKpfxltXH2siUGr00ESg+iUq2JdgXy9yivtOBI0tbby4OVdXIis1zGkiUP0iIqRGB9k1hfS9L07w07f28uwG+4raKaVcQxOB6rc0O6eQbs+1VD3925ocSmubnB2WUmqANBGofkuLCuJEVSO1Ta29nrcjr5K0qEAaW9p4+KNDQxSdUqq/NBGofuuYOXS0l+6h2qZWsoqquWTaWK6fl8RLW/I5fLJmqEJUSvWDJgLVb2l2zBzanV9Ju4HZ48K4+4J0Anw8+e17B4YqRKVUP2giUP2WFBGAp4f0mgg6xgdmJI4hPNCHOxaNZ01WCRsOlw5VmEopO2kiUP3m6+VJUnhAr4lgR14F6TFBhPp7A3DD2ckkhPnz4Lv7e9wARynlGpoI1ICkRQWSU2x7jKC93bAjt4JZSWGdx/y8PXlgaSYHi2p4Y3vBUIWplLKDJgI1IGlRQRwtrbP57f5IaS3Vja3MGhd2yvFLpsYxK2kMf1ydRV0fM46UUkNHE4EakLSoIJrb2imoOH3byo7xga4tArAsRvvpJZMormniqXWnb4uplHINTQRqQNKiey4+tyO3kjEB3qTa2Blt9rgwLpkWx5PrciiqanR6nEqpvmkiUAOSGmmdQmpjnGB7nmV8oKcKpg8syaSlzbByU65TY1RK2UcTgRqQsEAfIgJ9TmsRVNW3kF1cy6ykMT1emxgewLyUcN7be8LuDW6OldaxI69iUDErpWzTRKAGLC0q6LREsCPfOj7QbaC4u6VT4zhSUsdhO6qYAtz72m6+/sxmyrRmkVIOp4lADVhadOBpVUh35lbgITA9oecWAcDiyTGIWCqU9iW3rI5tuRXUN7e5ZJD5RFUDO7U1okYxpyYCEVkiIlkiki0iD/Ry3nIRMSIyx5nxKMdKiwqivK6Z8rrmzmPb8yqYGBdCoG/vu6BGB/txxrhwPthb1Of7vLWzEBE4Z3wkL3x+jJKaoWsVtLcbbv3ndlY8vYmaxpYhe1+lhpLTEoGIeAKPAUuBScAKEZlk47xg4G5gs7NiUc7RUXPoiLV7qK3dsCuv8rRpoz1ZOjWWg0U1va5QNsbw1s5CzkqN4NeXT6GlzfDEpzmDD95Ob+8qZHdBFY0t7bz/Rd9JS6mRyJktgrlAtjHmiDGmGXgZuMzGeb8GHgJ0LuEI0734XFZRDXXNbczuY3ygw5IpsQC9tgp25FWQW1bPFTPjSYkM5MqZ8azclMvJauf/utQ3t/L7D7KYGh9KamQgr+uKaDVKOTMRxAP5XZ4XWI91EpFZQKIx5t3eXkhEbhGRbSKyraSkxPGRqgGJD/PHx8ujc5ygY1aPvS2CuFB/ZiaN4f29PY8TvLGjED9vD5ZOjQPgzvMn0NZueHyt81sFT356hKLqRn7+1Uksn53AlmPl5JWdvoBOqZHOZYPFIuIB/Bm4p69zjTFPGWPmGGPmREVFOT84ZRdPDyE1MrBz/+IduRVEBvmSGO5v92tcPCWOvYXVNj9gm1rbeHfPCRZPjiXIOuaQFBHA1XMS+NfmPE5UNTjmRmw4UdXAk+tyuGRaHGckh3PlrHhE4I0d2ipQo48zE0EhkNjleYL1WIdgYAqwVkSOAWcC7+iA8cjSdQrpjrwKZo8bg4jthWS2dHQP2WoVrDlYTFVDC1fMPKUhye2LxmMwPLYmexCR9+73H2TRbiyL38DSepmfFskbOwpo1+qpapRxZiLYCkwQkRQR8QGuBd7p+KExpsoYE2mMSTbGJAObgGXGmG1OjEk5WFpUIHnl9ZyoauBYWb3d3UIdEsMDmBofyvs2xgne2FFIVLAv54yPPOV4QlgAXzsjkVe25tusdTRYu/IreWtnId85J4XE8IDO41fNTqCgooEtx8od/p5KuZLTEoExphW4A1gFHABeNcbsE5FficgyZ72vGlpp0UG0G3hzh6WxZ+9AcVdLp8ayK7+Swsovu3oq6ppZm1XMZdPH4uV5+q/p7YvGI4jDWwXGGH793/1EBvnyvUXjT/lZRxeVltFWo41TxwiMMe8ZY9KNMWnGmN9Yj/3cGPOOjXMXamtg5OmYOfTatny8PYUp8aH9fo2lUywDwV1nD/13z3Fa2gxXzkqweU1cqD/XzUvitW0FDh3A/e+eE2zPreC+xemd4xId/H08uWRqHO99cYL65tFfRtsYY3cJEDWy6cpiNSgp1gqjx8rqmTw2FD9vzwG9RmZsMB90GSd4Y0chmbHBTBob0uN1ty1Mw9NDeOSTw/0P3IbGljZ+9/5BJsWFcNXsRJvnLJ+dQF1zm10L4Ua6lZvzmPfbj2lobnN1KMrJNBGoQQn09WJsqB9g/7RRWy6eGse23ApOVjdypKSWXfmVpw0SdxcT4sfXzxzHmzsKyC8ffKvg2Q1HKaxs4GeXTsSzh8qpZySHkRQeMOpnD7W3G55ed4TimiY+PaRTtkc7TQRq0NKiLd1DAxkf6HDx1FiMgVX7inhrZyEeApf3kQgAbl6QiocIz244OuD3Biiva+bxtTlcOCmGs9MiezxPRFg+K4HPcspOGdMYbTZkl5JnTa6r9o3+1o+700SgBq1jnGDWuN4LzfVmfHQw46ODeHfPCd7aWcj88ZHEhPj1eV1sqB/LZozl1W35VNUPvBbQE5/mUNfcyo8WZ/R57pWz4jEG3hrFrYIXN+cSEejDsulj+ejASZpb210dknIiTQRq0K6fl8RPLs4kLtT+hWS2XDwlls1HyymoaODKWX23BjrcvCCV+uY2XtwysI1uiqoaeeGzY1wxM54JMcF9nt+xn8IbOwqH3WBqbVMrSx5ex/t2VHXtSVFVIx8dKObqOYksmz6WmsZWNh0pc2CUarjRRKAGbUJMMLecmzbo1+koIxHg48niybF2XzcxLoQFEyJ5fuMxmlr7P7D56CeHaTeGH1yQbvc1V81O4Ogw3Czn4IlqDhbVcN/rezhWevrucfZ4eWsebe2G6+Ymcc6ESAJ8PPlAu4dGNU0EatjIjA1m8tgQrpwVT4BP72Wsu7vl3FSKa5p4Z9fxfl2XW1bHK1vzWTE36ZTFY31ZOjUOf29PXt9e2PfJVu3thuWPf8Y9r+6mrsk50087Vnm3trdzx0s7+p0YW9vaeXlLPuemR5EUEYCftyeLMqJZve8kbbqietTSRKCGDRHh37fP51fLpvT72nPGR5IZG8zT64/0q7vm4Y8O4+Up3NFt8Vhfgny9WDollv/uOU5ji30ftl8UVrE9t4I3dhRw+WMbyS6u6dd72iOnpA5fLw/+79qZ7C2s5n/fO9iv6z85WExRdSPXz0vqPLZ4SiyltU3DrvWjHEcTgRpWvDw9etz0vjciwi3npnLoZK3d0x2zimp4e1chN56dQrQdA9PdXT4znprGVjZml9p1/pqsYkTgkRUzKa9rZtlfN/LO7v61YPqSU1xLSmQgiyfH8q35KTz/2bF+zfp5cXMeMSG+fCUzuvPYoowofDw9WOUGayfclSYCNWpcOm0ssSF+PL3evu0s/7Q6iyAfL249L3VA73dmagTBvl58uP+kXeevySphRuIYlk0fy7t3LWBSXAh3vbSTn/9774DGNmzJKantnM77wNJMpiWEct9ru+2qyZRXVs+6wyVce0bSKWU9gv28mT8+gg/2FQ27wXHlGJoI1Kjh4+XBjfOT2Zhdxt7Cql7P3ZVfyer9J7nl3FTGBPgM+P3Oy4jiowPFfVYkLa1tYk9BJYsyLN+0Y0P9eOmWM7l5QQr/+DyXa574fNAF9Jpa28grr++czuvj5cGjK2ZiDNz50k5a2nqfAvrS1jw8RFgxN+m0ny2ZEktBRQP7jlcPKkY1PGkiUKPKirlJBPp48kwfrYI/rDpIRKAPN52TMqj3u3BSDKW1TewqqOz1vHWHSjCGzkQA4O3pwU8vmcQTX5/NkZI67n1t96BiyS2rp91YKsJ2GBcRyP8un8rOvEr+uDqrx2ubWtt4dWs+X8mMJjb09G6yCybG4CG6uGy00kSgRpVQf2+unZvEf/ac4HgPK383ZpeyMbuM7y0af1phuf5amBGNl4f02T20JquEyCBfJtuonbRkSizXnzmO7bkVg6rr07FBUEeLoMOl08Zy/bwknvz0CCs35dpsGasfi2IAABfPSURBVKzad5KyumauP3OczdeOCPJlbkr4kCWCvYVV/L+39+pMpSGiiUCNOjfNTwbguY1flp2obWple245Kzfl8ot39jE21O+UmTEDFervzbzU8F4TQWtbO59mFbMwI6rHgfB5qeG0tJlBzczpmDqa2qVF0OH/XTqJWUlj+Nnbe1nw0Bqe+DSHqoYvV2K/uCmXxHB/FozvubzG4smxHDpZ2/k+ztLY0sbdL+/kn5tyySpy/MwqdbrBfR1SahhKCAvgkqlxvLQln7zyeg4W1ZDbpVR1iJ8Xf7h6+oAqpdpy4cQYfvGf/RwtreusxtrVzvxKqhtbT+kW6m7OuDA8BDYfKWN+Lx/GvckpqSN+jL/NNRh+3p68fuvZrD1UzDPrj/K79w/yyMeHuWZOIudlRLH5aDn3L8nsdcbW4smx/PI/+1m1r4jvLezfdNv++Osn2Z37YO89XtVrBVrlGJoI1Kh063lpfHKwmMMna5kyNpSrZyeQGRvCxLEhjA3169d2mn25YJIlEXy4v8jmCus1B4vx9BAWpPf8AR/s582U+FA2HR347mc5JbU2WwMdPDyE8zNjOD8zhn3Hq3h2/VFWbsrl+c+O4e0pXDPH9t4PHcaO8Wd6Qiir9jovEew7XsXjn+Zw5ax4Vu0tYl9hFcyxXRJcOY4mAjUqTRobwt5fLh6S90oIC2BiXAgf7j9pOxFklTBnXBghft69vs68lHBe+DyXxpa2frdWjDHkFNdytZ0fmpPHhvLnr83gR0syWbkpl+gQXyKCfPu87qLJsfxhVRbHKxsYO2ZwtaW6a21r50ev7yEswIefXzqJgvIG9uospSGhYwRKOcCFk2LYnltBWW3TKceLqho5cKKaRZk9dwt1mJcSQXNrO7vze5+BZMvJ6ibqmts61xDYKzbUj3sXZ/DNs5LtOn/JFEsNqNVOGDR+ev1R9h2v5sHLJzMmwIfJ8SHsP16tA8ZDQBOBUg5w0aQY2o2lRENXa7Msz3sbH+hwRko4IrB5AN1D2Z0zhnruGnKEtKggJkQHObwI3ZGSWv7y0SGWTolliXXr0iljQ2loaeNoqXMHp5UmAqUcYrJ17KH77KE1WcWMDfUjPabvb+qh/t5MjA1h89H+l3zumMkzPqp/LYKBWDIlli1Hy09r/QxUe7vh/jf24O/tyS8vm9x5vGP/672F2j3kbJoIlHIAEeGCSTGsP1zaWYSuubWdDYdLWZgZbffg9LzUcLbnVvR7I5icklqCfb2ICu67n3+wFmZE0W5ge65jitC9uDmXrccq+H+XTiI6+MvFbGlRgfh6efS5SlwNniYCpRzkgokxNLS0seGwpQjd1mPl1DW3cb4d3UId5qWE09jSzheF/RsnyCmpJTU6yKGzoXqSEWuZzumIOf6FlQ387v2DnJsexfJumxF5eXowMS6EvcftSwQf7C3iRNXo3T7UmTQRKOUg3YvQrTlYjI+nB2ePj7D7NeamWM7ddKR/4wQ5xXVOHx/oEOTrRWK4PwdPDjwRtLUbXt9ewFWPf4YBfnvFFJtJbPLYEPYVVvdZy6m4ppFbV27n/je+GHBM7kwTgVIO0lGE7uODJ2lvN6zJKmZeani/NtkJD/QhPSaoXwPGtU2tFFU3nlZawpkyYkIG3CJYf7iESx/dwL2v7SY62Jd/fnseCWG2NwWaEh9KTVMr+X0U5Otoha07VMJm3Vaz3zQRKOVAliJ0zfxnz3FySursmi3U3byUCLYfK6e1j2qhHY6U2K4x5EyZscEcLa3rV/nsAyeq+ebft/CNZ7dQ09jCIytm8tb35jN7XFiP10wZa9+A8bpDJUQE+hAd7MsfV2dpuex+0kSglAN1FKH7zbsHAOxaP9DdvNRw6prb7F5M1TljKHpouoYAMmKDaWs3ndNW+/LwR4e4+JH17M6v5GeXTOTje85j2fSxfW5ClB4bhJeH9DpO0N5u2JBdyoIJkdz5lQlsPVbBWjs3JxopmlrbWP74Z3xk594X/aWJQCkH6ihCV1zTRHJEgM3aQ32ZmxIOwBY7p5HmFNfh5SGMixi6RJAZGwzYN2Dc2tbOk58e4dwJUXx630K+syAVXy/7Vk77enmSHhPc68yh/SeqKa1t5tz0KL42J5GEMH/+uCqrz3GFkeQ/u0+wPbfCYfWxutNEoJSDXTgxBhhYawAgOtiP1KhANts5YJxTUktSRADenkP3v3NyZCA+nh52JYLskloaWtq4fObYAW0CNMW6wrin7p711vGBc8ZH4uPlwQ8uSGff8WqHL3pzFWMMf99wlPSYIOb3Y+JBf2giUMrBLp4aR1pUIFfMjO/75B7MS4lgy7Fyu8or5JTUDun4AFg21UmLDuKgHYmgo2TG9IQxA3qvKfGhlNU1U1TdaPPn6w6VkBkb3Lnv9OUz4xkfHcSfVmf1+u9X29RKfvngdoUbCpuPlrP/RDXfmp/itOnBmgiUcrDoED8+vmch0wb4wQdwZmo4NY2tHDjR+zhBa1s7x0rrhzwRAEyMDbarRbArv4oQPy+SB9h1NbmXAeP65la25ZZzXnpU5zFPD+GeC9PJKanjrZ2FNl9zd34lSx5ex8I/ruX3Hxx02J7RzvD3DUcJC/Dm8kF8seiLJgKlhqGOcYK+ppEWVDTQ3NY+ZGsIusqIDaaoupHK+uZez9udX8n0xDF9Dgz3ZGJcMB6CzXGCTUfKaGkzLJgQdcrxJVNimRofysMfHTpllbYxhuc3HuWqJz7DGLh0Whx/W5vDpY9sGFCxP2fLK6vnwwMnuX7eOKeND4AmAqWGpbhQf5LCA/qcE98xY6i/VUcdIcM6YNxb91BDcxtZJ2sG3C0EEODjRVpUEPtszBxad6gUP28P5iSfOgVVRLjnonQKKhp4ZWseAFUNLdy2cge/+M9+zkuP4t27zuH/rp3J8zedQW1TK1f8bSO/e/9gZ4mQ4eD5z47hKcI3zrK9haijaCJQapialxLOlmPlvc5+6UwEkUOfCDLtKDWx73gVbe2G6YkDTwRgGSew1TW07nAJ81IibH5bPi89irnJ4Tz6STZbjpbz1Uc38NGBk/z04ok8/c05nQPXCzOiWfWDc7lmTiJPfJrDpY9uYOcgtgx1lJrGFl7dls+l0+KICfHr+4JB0ESg1DA1LzWCyvoWDhX3/EGbU1xHZJAvoQG9b3rjDDEhvoT6e/faItjVMVCcGDqo95o8NoSi6kZKar6seFpQUc+RkjoWTLC985uIcO/iDIprmrjmyc9pbWvnle+exc3npp426Bri583vlk/jhW/Npb6plauf+LzP8Rl77Dte1flv0F+vby+gtqmVm+anDDqOvmgiUGqYmtcxTtDLNFLLjKGhHx8AywdtRmwwWUU9f2Duyq8kfoz/KVVFB6KjJHXX7qGOaaNdB4q7m5sSzrVnJHLJ1DjevWtBr6uYO17r3bsWEODjye8/ODiomAHueXU333p+Kw3N/etuams3PP/ZMWaPCxt0a8oemgiUGqYSwwOIH+Pf6/4EOSW1Lhkf6JAZG8yhk7U9zvHfXVA56NYA0LmB/b4uq63XHy4hNsSP8X3c/++WT+Ox62cRFmjfGoawQB9uWzieNVmDq1t0vLKBg0U1lNc1d45T2OuTg8XkltXzrSFoDYAmAqWGtXkp4XyWU2ZzZk55XTMV9S0umTraISM2mNqmVgoqTi//XFbbRH55w6AGijuE+HmTHBHQOXOord2w4bClrIQz5tbfeHYyMSG+PPTBwQHXLVqbZSlzkRjuz9Prj9JiZ+0osEwZHRvqx+LJMQN67/5yaiIQkSUikiUi2SLygI2f3yoiX4jILhHZICKTnBmPUiPNjfOTqW9q486Xdp5WhK5zoNhFXUPQe6mJPQWWD21HdW1Mjg/trDm0u6CS6sZWzu2lW2gw/H08+f4F6ezIqzxt1zl7rc0qJn6MP79cNpnCygb+veu4XdftP17N50fKuOHsZLyGaLW4095FRDyBx4ClwCRghY0P+n8ZY6YaY2YAvwf+7Kx4lBqJpiWM4cHLp7D+cCm/X5V1ys9yioe+6mh36THWRGBjb4Jd+ZV4CEyNH3zXEFgqkeaXN1BV38L6Q6WIWMpKOMvVsxNIjQrkD6t6X6FsS3NrOxuzSzkvI4pFGdFkxgbzxKc5dtU/em7jUfy9Pbn2jKSBht5vzkw3c4FsY8wRY0wz8DJwWdcTjDFdR5kCgdFTJUopB7nmjERuOGscT607wttdVsrmlNTi6+VB/Bh/l8UW7OdN/Bh/mzOHdhdUMiE6mEBf+/dj6M2U+I5xgirWHS5hWnyo3f3+A+Hl6cF9F2VwuLiWN3cU9Ovabdbd6RZlWLYpvW1hGtnFtazuo3VRUtPEv3cf56rZCUM6E8yZiSAeyO/yvMB67BQicruI5GBpEdxl64VE5BYR2SYi20pKRld5WaXs8bNLJzE3JZz739jT2U+eXVxLalTQgFfsOkqmjZlDxhjrimLHtAbgy1ITn+WUsSu/8rTVxM6wZEos0xNC+cuHh/q10GxNlnV3ujRLkbhLpsaRFB7A42uzexxzaGpt4/Z/7QADN81PdkT4dnP5YLEx5jFjTBpwP/CzHs55yhgzxxgzJyrK+f/xlRpuvD09+Nv1s4gI9OGWf2yjtLaJnJKh256yN5lxwRwpqTullEN+eQMV9S0OnfoYHuhD/Bh/Vm7Opa3d9Lh+wJFEhPuXZHK8qpGVm3Ltvm5tVglzU8I7W0Nenh7cel4auwuq+Czn9JlIxhgeeOMLthwt5w9XTyN1iLv7nJkICoHELs8TrMd68jJwuRPjUWpEiwzy5clvzKGsrpnbVm4nv8I1xea6y4gNobXddA5eA+wqGFzF0Z5MHhtCZX0LgT6ezOpjTYCjnD0+kgUTIvnrmmyqG1v6PL+gop7DxbUszDj1S+vy2fFEB/vyt7XZp13zl48O89bOQu69KJ3LZjivuFxPnJkItgITRCRFRHyAa4F3up4gIhO6PL0EOOzEeJQa8aYmhPLQ8mlsPVaBMa6pMdSdrZlDu/Mr8fXy6KxH5CgdC8vOSosc0v0X7l+SSWV9C0+vO9LnuR3TRhd226bU18uT7yxIYWN22SkF7l7bls8jHx/mmjkJ3L5ovGMDt5PT/iWNMa3AHcAq4ADwqjFmn4j8SkSWWU+7Q0T2icgu4IfADc6KR6nR4vKZ8dy8wLLQaFKcYz9oByIlMhBvT+FAl3GC3fmVTIkPdfiHdceA8Xnpzu8WOvV9Q7l0WhzPrD9KcY3tfRE6rM0qJiHM32a33XXzxhHq793ZKvgsu5Qfv/kF54yP5DdXTHXafgN9cWpKNca8Z4xJN8akGWN+Yz32c2PMO9bHdxtjJhtjZhhjFhlj9jkzHqVGix8vnchHPzyX8dGuTwTenh6kRQV1tgha2trZe7yKGU4ojTB/fCQ/vDDdqbX5e3LvRRm0trfzp1WHejynqbWNjdllnbOFugvy9eKGs8axat9J3v/iBN9duZ3UqED+9vVZQ9rC6c7lg8VKqf7z8JBhkQQ6ZHbZpObQyRoaW9qdUiPH18uTu74ygWC/oS+ylxwZyE3zU3h1ez57CmwXkttytJyGljYWZfY8qeXG+Sn4e3ty24s78PP25O83nkGIC+6nK00ESqlBy4gN4URVI1X1LZ3VNmc4eKB4OLjz/PFEBPryi3f22ZwGujarBB8vD85K7bnrKjzQhxvOTibAx5Nnb5hDQliAM0O2iyYCpdSgdQ4Yn6xhd34lYQHeJIa7bqGbswT7efOjJRnsyKvk7V2nT4Jck1XMmakR+Pv0vpvY/Usy2PyTrwxqO1NH0kSglBq0jM6ZQ9Xszq9ieuIYlw18OttVsxKYnhDK794/SF1Ta+fxvDLL/ggL7ah/JCIu6d7qiSYCpdSgxYX6EeznxY68Sg4VD25ryuHOw0P4n2WTOVndxGNrvlwTsPZQMQCLMqN7unTY0kSglBo0ESEzNpgP9hZhDE6ZMTSczEoK48qZ8Tyz/ii5ZXUArDlYzLiIAFIiXb/au780ESilHCIjNpgGaz2eaQmOqzE0XN2/NBMvT+HBdw/Q2NLG50cs00ZHIk0ESimH6NjMPjHcn4ggXxdH43wxIX7cef4EPtx/kj+uyqKxpZ3zMkZmLTRNBEoph+iYOTSaxwe6+9Y5ySRHBPDMhqP4enlwVmqEq0MaEE0ESimHyIgNJsjXi/lO3CxmuPH18uRnl1j22zorLQI/796njQ5XjtkxQinl9oL9vNn4wPkEO2gjmpHiKxOjuW9xBvNSwl0dyoC5138xpZRThfoPn7nxQ0VEXFY11FG0a0gppdycJgKllHJzmgiUUsrNaSJQSik3p4lAKaXcnCYCpZRyc5oIlFLKzWkiUEopNye2tlsbzkSkBMgd4OWRQKkDwxkp3PW+wX3vXe/bvdhz3+OMMTar4o24RDAYIrLNGDPH1XEMNXe9b3Dfe9f7di+DvW/tGlJKKTeniUAppdycuyWCp1wdgIu4632D+9673rd7GdR9u9UYgVJKqdO5W4tAKaVUN5oIlFLKzblNIhCRJSKSJSLZIvKAq+NxFhH5u4gUi8jeLsfCReRDETls/TvMlTE6g4gkisgaEdkvIvtE5G7r8VF97yLiJyJbRGS39b5/aT2eIiKbrb/vr4iIj6tjdQYR8RSRnSLyX+vzUX/fInJMRL4QkV0iss16bFC/526RCETEE3gMWApMAlaIyCTXRuU0zwNLuh17APjYGDMB+Nj6fLRpBe4xxkwCzgRut/43Hu333gScb4yZDswAlojImcBDwF+MMeOBCuDbLozRme4GDnR57i73vcgYM6PL2oFB/Z67RSIA5gLZxpgjxphm4GXgMhfH5BTGmHVAebfDlwEvWB+/AFw+pEENAWPMCWPMDuvjGiwfDvGM8ns3FrXWp97WPwY4H3jdenzU3TeAiCQAlwDPWJ8LbnDfPRjU77m7JIJ4IL/L8wLrMXcRY4w5YX1cBMS4MhhnE5FkYCawGTe4d2v3yC6gGPgQyAEqjTGt1lNG6+/7w8CPgHbr8wjc474NsFpEtovILdZjg/o9183r3YwxxojIqJ0zLCJBwBvA940x1ZYviRaj9d6NMW3ADBEZA7wFZLo4JKcTkUuBYmPMdhFZ6Op4htg5xphCEYkGPhSRg11/OJDfc3dpERQCiV2eJ1iPuYuTIhIHYP272MXxOIWIeGNJAi8aY960HnaLewcwxlQCa4CzgDEi0vFFbzT+vs8HlonIMSxdvecD/8fov2+MMYXWv4uxJP65DPL33F0SwVZggnVGgQ9wLfCOi2MaSu8AN1gf3wD824WxOIW1f/hZ4IAx5s9dfjSq711EoqwtAUTEH7gQy/jIGuAq62mj7r6NMT82xiQYY5Kx/P/8iTHmekb5fYtIoIgEdzwGLgL2Msjfc7dZWSwiF2PpU/QE/m6M+Y2LQ3IKEXkJWIilLO1J4H+At4FXgSQsJbyvMcZ0H1Ae0UTkHGA98AVf9hn/BMs4wai9dxGZhmVw0BPLF7tXjTG/EpFULN+Uw4GdwNeNMU2ui9R5rF1D9xpjLh3t9229v7esT72AfxljfiMiEQzi99xtEoFSSinb3KVrSCmlVA80ESillJvTRKCUUm5OE4FSSrk5TQRKKeXmNBEo1Y2ItFkrO3b8cVihOhFJ7loZVqnhQEtMKHW6BmPMDFcHodRQ0RaBUnay1oH/vbUW/BYRGW89niwin4jIHhH5WESSrMdjROQt614Bu0XkbOtLeYrI09b9A1ZbVwQr5TKaCJQ6nX+3rqGvdflZlTFmKvBXLCvVAR4FXjDGTANeBB6xHn8E+NS6V8AsYJ/1+ATgMWPMZKASWO7k+1GqV7qyWKluRKTWGBNk4/gxLJvAHLEWuCsyxkSISCkQZ4xpsR4/YYyJFJESIKFriQNriewPrRuIICL3A97GmAedf2dK2aYtAqX6x/TwuD+61r5pQ8fqlItpIlCqf77W5e/PrY8/w1IBE+B6LMXvwLJl4G3QuXlM6FAFqVR/6DcRpU7nb93xq8MHxpiOKaRhIrIHy7f6FdZjdwLPich9QAlwk/X43cBTIvJtLN/8bwNOoNQwo2MEStnJOkYwxxhT6upYlHIk7RpSSik3py0CpZRyc9oiUEopN6eJQCml3JwmAqWUcnOaCJRSys1pIlBKKTf3/wHcqGLytqO3nwAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "glove = torchtext.vocab.GloVe(name=\"6B\", dim=300)\n",
        "\n",
        "model = Transformer(300).cuda()\n",
        "train_network(model, train_data, valid_data, batch_size = 512, num_epochs=50, learning_rate=2e-4)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "3e-1kafPLBH7",
        "outputId": "d36a3ef4-4d70-403a-ecd1-54af33e145b8"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|█████████▉| 399999/400000 [00:38<00:00, 10466.21it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1; Loss 0.582342; Train Acc 0.725553; Val Acc 0.714896\n",
            "Epoch 2; Loss 0.367320; Train Acc 0.841899; Val Acc 0.827549\n",
            "Epoch 3; Loss 0.289710; Train Acc 0.887045; Val Acc 0.869194\n",
            "Epoch 4; Loss 0.246769; Train Acc 0.907923; Val Acc 0.879338\n",
            "Epoch 5; Loss 0.246298; Train Acc 0.930942; Val Acc 0.897491\n",
            "Epoch 6; Loss 0.154612; Train Acc 0.948430; Val Acc 0.908169\n",
            "Epoch 7; Loss 0.106439; Train Acc 0.962170; Val Acc 0.910304\n",
            "Epoch 8; Loss 0.099253; Train Acc 0.977159; Val Acc 0.914576\n",
            "Epoch 9; Loss 0.085501; Train Acc 0.987866; Val Acc 0.914042\n",
            "Epoch 10; Loss 0.053384; Train Acc 0.991970; Val Acc 0.914042\n",
            "Epoch 11; Loss 0.026607; Train Acc 0.995360; Val Acc 0.917779\n",
            "Epoch 12; Loss 0.020731; Train Acc 0.996788; Val Acc 0.915643\n",
            "Epoch 13; Loss 0.008345; Train Acc 0.996788; Val Acc 0.911372\n",
            "Epoch 14; Loss 0.006701; Train Acc 0.997859; Val Acc 0.915643\n",
            "Epoch 15; Loss 0.002726; Train Acc 0.999465; Val Acc 0.918313\n",
            "Epoch 16; Loss 0.004130; Train Acc 0.999108; Val Acc 0.910838\n",
            "Epoch 17; Loss 0.024232; Train Acc 0.999286; Val Acc 0.910838\n",
            "Epoch 18; Loss 0.007294; Train Acc 0.993219; Val Acc 0.910838\n",
            "Epoch 19; Loss 0.037556; Train Acc 0.997859; Val Acc 0.918847\n",
            "Epoch 20; Loss 0.011762; Train Acc 0.998929; Val Acc 0.918847\n",
            "Epoch 21; Loss 0.003040; Train Acc 0.998929; Val Acc 0.917245\n",
            "Epoch 22; Loss 0.005959; Train Acc 0.998216; Val Acc 0.913508\n",
            "Epoch 23; Loss 0.006662; Train Acc 0.999286; Val Acc 0.916711\n",
            "Epoch 24; Loss 0.003491; Train Acc 0.999286; Val Acc 0.918313\n",
            "Epoch 25; Loss 0.002382; Train Acc 0.999108; Val Acc 0.919381\n",
            "Epoch 26; Loss 0.010010; Train Acc 0.999108; Val Acc 0.914576\n",
            "Epoch 27; Loss 0.001730; Train Acc 0.998929; Val Acc 0.919915\n",
            "Epoch 28; Loss 0.012580; Train Acc 0.999286; Val Acc 0.917779\n",
            "Epoch 29; Loss 0.000906; Train Acc 0.999465; Val Acc 0.923118\n",
            "Epoch 30; Loss 0.008590; Train Acc 0.999286; Val Acc 0.922584\n",
            "Epoch 31; Loss 0.006762; Train Acc 0.998751; Val Acc 0.922050\n",
            "Epoch 32; Loss 0.002492; Train Acc 0.999465; Val Acc 0.921516\n",
            "Epoch 33; Loss 0.000357; Train Acc 0.999822; Val Acc 0.916177\n",
            "Epoch 34; Loss 0.000661; Train Acc 1.000000; Val Acc 0.913508\n",
            "Epoch 35; Loss 0.002054; Train Acc 0.999822; Val Acc 0.919915\n",
            "Epoch 36; Loss 0.002079; Train Acc 0.998394; Val Acc 0.922050\n",
            "Epoch 37; Loss 0.003093; Train Acc 0.999465; Val Acc 0.916177\n",
            "Epoch 38; Loss 0.001359; Train Acc 0.999286; Val Acc 0.920448\n",
            "Epoch 39; Loss 0.015536; Train Acc 0.999643; Val Acc 0.915643\n",
            "Epoch 40; Loss 0.010845; Train Acc 0.998751; Val Acc 0.912440\n",
            "Epoch 41; Loss 0.017783; Train Acc 0.999286; Val Acc 0.916711\n",
            "Epoch 42; Loss 0.030678; Train Acc 0.999822; Val Acc 0.912974\n",
            "Epoch 43; Loss 0.036529; Train Acc 0.997502; Val Acc 0.914042\n",
            "Epoch 44; Loss 0.006387; Train Acc 0.999465; Val Acc 0.914576\n",
            "Epoch 45; Loss 0.021423; Train Acc 0.997502; Val Acc 0.910304\n",
            "Epoch 46; Loss 0.000785; Train Acc 0.999108; Val Acc 0.919381\n",
            "Epoch 47; Loss 0.000868; Train Acc 0.999643; Val Acc 0.917245\n",
            "Epoch 48; Loss 0.006293; Train Acc 0.999108; Val Acc 0.920982\n",
            "Epoch 49; Loss 0.001005; Train Acc 0.999465; Val Acc 0.917245\n",
            "Epoch 50; Loss 0.004505; Train Acc 0.999465; Val Acc 0.918847\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "glove = torchtext.vocab.GloVe(name=\"6B\", dim=50)\n",
        "\n",
        "model = Transformer(50).cuda()\n",
        "train_network(model, train_data, valid_data, batch_size = 512, num_epochs=100, learning_rate=2e-3)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "4tNVR9NyLEdg",
        "outputId": "35ab3fd4-1fe6-4605-d6c6-d2093895af1c"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1; Loss 0.744703; Train Acc 0.508922; Val Acc 0.506140\n",
            "Epoch 2; Loss 0.699941; Train Acc 0.491078; Val Acc 0.493860\n",
            "Epoch 3; Loss 0.695198; Train Acc 0.508922; Val Acc 0.506140\n",
            "Epoch 4; Loss 0.689438; Train Acc 0.742862; Val Acc 0.725040\n",
            "Epoch 5; Loss 0.677176; Train Acc 0.720914; Val Acc 0.730913\n",
            "Epoch 6; Loss 0.487524; Train Acc 0.812099; Val Acc 0.797117\n",
            "Epoch 7; Loss 0.359105; Train Acc 0.859565; Val Acc 0.847304\n",
            "Epoch 8; Loss 0.361208; Train Acc 0.881513; Val Acc 0.852109\n",
            "Epoch 9; Loss 0.266175; Train Acc 0.899893; Val Acc 0.864389\n",
            "Epoch 10; Loss 0.325049; Train Acc 0.918808; Val Acc 0.877202\n",
            "Epoch 11; Loss 0.263693; Train Acc 0.912027; Val Acc 0.872931\n",
            "Epoch 12; Loss 0.195186; Train Acc 0.920949; Val Acc 0.861185\n",
            "Epoch 13; Loss 0.215853; Train Acc 0.938615; Val Acc 0.878270\n",
            "Epoch 14; Loss 0.173559; Train Acc 0.948965; Val Acc 0.880406\n",
            "Epoch 15; Loss 0.160708; Train Acc 0.955032; Val Acc 0.876135\n",
            "Epoch 16; Loss 0.158647; Train Acc 0.964311; Val Acc 0.879338\n",
            "Epoch 17; Loss 0.147644; Train Acc 0.956638; Val Acc 0.873465\n",
            "Epoch 18; Loss 0.134871; Train Acc 0.968951; Val Acc 0.883075\n",
            "Epoch 19; Loss 0.132524; Train Acc 0.971627; Val Acc 0.888414\n",
            "Epoch 20; Loss 0.119767; Train Acc 0.972698; Val Acc 0.882007\n",
            "Epoch 21; Loss 0.098892; Train Acc 0.965560; Val Acc 0.880940\n",
            "Epoch 22; Loss 0.077401; Train Acc 0.968594; Val Acc 0.878804\n",
            "Epoch 23; Loss 0.060556; Train Acc 0.982869; Val Acc 0.885745\n",
            "Epoch 24; Loss 0.129224; Train Acc 0.981620; Val Acc 0.888414\n",
            "Epoch 25; Loss 0.066707; Train Acc 0.958244; Val Acc 0.874533\n",
            "Epoch 26; Loss 0.074685; Train Acc 0.984118; Val Acc 0.883075\n",
            "Epoch 27; Loss 0.069311; Train Acc 0.983940; Val Acc 0.883075\n",
            "Epoch 28; Loss 0.039092; Train Acc 0.988401; Val Acc 0.883075\n",
            "Epoch 29; Loss 0.083804; Train Acc 0.989293; Val Acc 0.876668\n",
            "Epoch 30; Loss 0.068979; Train Acc 0.988580; Val Acc 0.875601\n",
            "Epoch 31; Loss 0.034758; Train Acc 0.988044; Val Acc 0.876668\n",
            "Epoch 32; Loss 0.064878; Train Acc 0.991970; Val Acc 0.883075\n",
            "Epoch 33; Loss 0.028540; Train Acc 0.980371; Val Acc 0.877202\n",
            "Epoch 34; Loss 0.035088; Train Acc 0.990007; Val Acc 0.883075\n",
            "Epoch 35; Loss 0.063860; Train Acc 0.982869; Val Acc 0.866524\n",
            "Epoch 36; Loss 0.038117; Train Acc 0.991970; Val Acc 0.879338\n",
            "Epoch 37; Loss 0.012270; Train Acc 0.992148; Val Acc 0.887347\n",
            "Epoch 38; Loss 0.047359; Train Acc 0.993041; Val Acc 0.882007\n",
            "Epoch 39; Loss 0.013694; Train Acc 0.993041; Val Acc 0.887347\n",
            "Epoch 40; Loss 0.036531; Train Acc 0.993398; Val Acc 0.877736\n",
            "Epoch 41; Loss 0.007078; Train Acc 0.993398; Val Acc 0.883075\n",
            "Epoch 42; Loss 0.037576; Train Acc 0.994468; Val Acc 0.891618\n",
            "Epoch 43; Loss 0.014309; Train Acc 0.995717; Val Acc 0.894821\n",
            "Epoch 44; Loss 0.028014; Train Acc 0.995717; Val Acc 0.887347\n",
            "Epoch 45; Loss 0.057386; Train Acc 0.995360; Val Acc 0.890016\n",
            "Epoch 46; Loss 0.024180; Train Acc 0.995004; Val Acc 0.884143\n",
            "Epoch 47; Loss 0.006553; Train Acc 0.996074; Val Acc 0.884677\n",
            "Epoch 48; Loss 0.035639; Train Acc 0.995539; Val Acc 0.884143\n",
            "Epoch 49; Loss 0.040968; Train Acc 0.996788; Val Acc 0.890550\n",
            "Epoch 50; Loss 0.011688; Train Acc 0.996074; Val Acc 0.894287\n",
            "Epoch 51; Loss 0.040929; Train Acc 0.995717; Val Acc 0.891084\n",
            "Epoch 52; Loss 0.028354; Train Acc 0.996253; Val Acc 0.888414\n",
            "Epoch 53; Loss 0.031999; Train Acc 0.993933; Val Acc 0.888948\n",
            "Epoch 54; Loss 0.007599; Train Acc 0.992505; Val Acc 0.886279\n",
            "Epoch 55; Loss 0.033189; Train Acc 0.995360; Val Acc 0.888948\n",
            "Epoch 56; Loss 0.018604; Train Acc 0.995896; Val Acc 0.890016\n",
            "Epoch 57; Loss 0.050766; Train Acc 0.990186; Val Acc 0.893753\n",
            "Epoch 58; Loss 0.046367; Train Acc 0.988401; Val Acc 0.889482\n",
            "Epoch 59; Loss 0.031218; Train Acc 0.991970; Val Acc 0.887880\n",
            "Epoch 60; Loss 0.024348; Train Acc 0.994825; Val Acc 0.884677\n",
            "Epoch 61; Loss 0.029289; Train Acc 0.991078; Val Acc 0.889482\n",
            "Epoch 62; Loss 0.047218; Train Acc 0.994111; Val Acc 0.896423\n",
            "Epoch 63; Loss 0.019624; Train Acc 0.995182; Val Acc 0.892152\n",
            "Epoch 64; Loss 0.025083; Train Acc 0.994647; Val Acc 0.895355\n",
            "Epoch 65; Loss 0.027081; Train Acc 0.996966; Val Acc 0.894287\n",
            "Epoch 66; Loss 0.025950; Train Acc 0.996253; Val Acc 0.893753\n",
            "Epoch 67; Loss 0.026907; Train Acc 0.992505; Val Acc 0.896423\n",
            "Epoch 68; Loss 0.008701; Train Acc 0.995539; Val Acc 0.890550\n",
            "Epoch 69; Loss 0.005402; Train Acc 0.997323; Val Acc 0.890550\n",
            "Epoch 70; Loss 0.003762; Train Acc 0.997145; Val Acc 0.896957\n",
            "Epoch 71; Loss 0.014777; Train Acc 0.997859; Val Acc 0.893219\n",
            "Epoch 72; Loss 0.013251; Train Acc 0.997502; Val Acc 0.892686\n",
            "Epoch 73; Loss 0.022020; Train Acc 0.992862; Val Acc 0.893219\n",
            "Epoch 74; Loss 0.007008; Train Acc 0.992862; Val Acc 0.890016\n",
            "Epoch 75; Loss 0.012299; Train Acc 0.993754; Val Acc 0.891084\n",
            "Epoch 76; Loss 0.006006; Train Acc 0.996966; Val Acc 0.887347\n",
            "Epoch 77; Loss 0.010009; Train Acc 0.997323; Val Acc 0.890016\n",
            "Epoch 78; Loss 0.038835; Train Acc 0.997680; Val Acc 0.887347\n",
            "Epoch 79; Loss 0.002188; Train Acc 0.997680; Val Acc 0.888414\n",
            "Epoch 80; Loss 0.008160; Train Acc 0.997680; Val Acc 0.888414\n",
            "Epoch 81; Loss 0.006122; Train Acc 0.997502; Val Acc 0.892152\n",
            "Epoch 82; Loss 0.026764; Train Acc 0.997323; Val Acc 0.886279\n",
            "Epoch 83; Loss 0.049123; Train Acc 0.997502; Val Acc 0.887880\n",
            "Epoch 84; Loss 0.010585; Train Acc 0.997323; Val Acc 0.891084\n",
            "Epoch 85; Loss 0.005365; Train Acc 0.997502; Val Acc 0.890550\n",
            "Epoch 86; Loss 0.004950; Train Acc 0.998037; Val Acc 0.892152\n",
            "Epoch 87; Loss 0.011790; Train Acc 0.997680; Val Acc 0.887880\n",
            "Epoch 88; Loss 0.020071; Train Acc 0.996788; Val Acc 0.893753\n",
            "Epoch 89; Loss 0.023962; Train Acc 0.998572; Val Acc 0.890016\n",
            "Epoch 90; Loss 0.006021; Train Acc 0.997680; Val Acc 0.892686\n",
            "Epoch 91; Loss 0.002954; Train Acc 0.997145; Val Acc 0.898025\n",
            "Epoch 92; Loss 0.010582; Train Acc 0.998216; Val Acc 0.893753\n",
            "Epoch 93; Loss 0.010617; Train Acc 0.998572; Val Acc 0.889482\n",
            "Epoch 94; Loss 0.008543; Train Acc 0.997859; Val Acc 0.891618\n",
            "Epoch 95; Loss 0.013520; Train Acc 0.998572; Val Acc 0.887347\n",
            "Epoch 96; Loss 0.015252; Train Acc 0.996253; Val Acc 0.894821\n",
            "Epoch 97; Loss 0.032288; Train Acc 0.996788; Val Acc 0.896423\n",
            "Epoch 98; Loss 0.003073; Train Acc 0.990542; Val Acc 0.883609\n",
            "Epoch 99; Loss 0.046436; Train Acc 0.997145; Val Acc 0.890550\n",
            "Epoch 100; Loss 0.026918; Train Acc 0.996610; Val Acc 0.887880\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "glove = torchtext.vocab.GloVe(name=\"6B\", dim=300)\n",
        "\n",
        "model = Transformer(300).cuda()\n",
        "train_network(model, train_data, valid_data, batch_size = 512, num_epochs=100, learning_rate=2e-4)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "xtIK0y3BOFUY",
        "outputId": "be660079-8bfc-4f2e-b928-7c2ac50d0a50"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1; Loss 0.490149; Train Acc 0.809422; Val Acc 0.815804\n",
            "Epoch 2; Loss 0.409235; Train Acc 0.859208; Val Acc 0.853711\n",
            "Epoch 3; Loss 0.253351; Train Acc 0.887580; Val Acc 0.876668\n",
            "Epoch 4; Loss 0.209999; Train Acc 0.913990; Val Acc 0.896957\n",
            "Epoch 5; Loss 0.170127; Train Acc 0.933619; Val Acc 0.900694\n",
            "Epoch 6; Loss 0.159220; Train Acc 0.948073; Val Acc 0.907635\n",
            "Epoch 7; Loss 0.109413; Train Acc 0.967166; Val Acc 0.918313\n",
            "Epoch 8; Loss 0.105485; Train Acc 0.978051; Val Acc 0.920448\n",
            "Epoch 9; Loss 0.096615; Train Acc 0.986617; Val Acc 0.917779\n",
            "Epoch 10; Loss 0.024777; Train Acc 0.991792; Val Acc 0.918313\n",
            "Epoch 11; Loss 0.015384; Train Acc 0.998037; Val Acc 0.917245\n",
            "Epoch 12; Loss 0.011698; Train Acc 0.998394; Val Acc 0.916711\n",
            "Epoch 13; Loss 0.004004; Train Acc 0.999465; Val Acc 0.917779\n",
            "Epoch 14; Loss 0.004631; Train Acc 0.999822; Val Acc 0.914576\n",
            "Epoch 15; Loss 0.002741; Train Acc 0.999108; Val Acc 0.916177\n",
            "Epoch 16; Loss 0.004893; Train Acc 0.998929; Val Acc 0.918313\n",
            "Epoch 17; Loss 0.003036; Train Acc 0.999108; Val Acc 0.922050\n",
            "Epoch 18; Loss 0.007009; Train Acc 0.996966; Val Acc 0.920982\n",
            "Epoch 19; Loss 0.015879; Train Acc 0.998394; Val Acc 0.916711\n",
            "Epoch 20; Loss 0.005161; Train Acc 0.997680; Val Acc 0.915109\n",
            "Epoch 21; Loss 0.015959; Train Acc 0.999108; Val Acc 0.919381\n",
            "Epoch 22; Loss 0.013455; Train Acc 0.998572; Val Acc 0.920448\n",
            "Epoch 23; Loss 0.026641; Train Acc 0.996253; Val Acc 0.915643\n",
            "Epoch 24; Loss 0.012780; Train Acc 0.996966; Val Acc 0.920448\n",
            "Epoch 25; Loss 0.008224; Train Acc 0.997859; Val Acc 0.919381\n",
            "Epoch 26; Loss 0.007812; Train Acc 0.999465; Val Acc 0.923652\n",
            "Epoch 27; Loss 0.001899; Train Acc 0.999286; Val Acc 0.922050\n",
            "Epoch 28; Loss 0.004505; Train Acc 0.998929; Val Acc 0.923652\n",
            "Epoch 29; Loss 0.000807; Train Acc 0.998929; Val Acc 0.920448\n",
            "Epoch 30; Loss 0.014531; Train Acc 0.999465; Val Acc 0.925254\n",
            "Epoch 31; Loss 0.006640; Train Acc 0.999286; Val Acc 0.926321\n",
            "Epoch 32; Loss 0.005180; Train Acc 0.998929; Val Acc 0.922050\n",
            "Epoch 33; Loss 0.001983; Train Acc 0.999108; Val Acc 0.923118\n",
            "Epoch 34; Loss 0.015482; Train Acc 0.999108; Val Acc 0.921516\n",
            "Epoch 35; Loss 0.020144; Train Acc 0.998572; Val Acc 0.917779\n",
            "Epoch 36; Loss 0.006975; Train Acc 0.998572; Val Acc 0.920448\n",
            "Epoch 37; Loss 0.006968; Train Acc 0.999108; Val Acc 0.920448\n",
            "Epoch 38; Loss 0.002862; Train Acc 0.998751; Val Acc 0.919381\n",
            "Epoch 39; Loss 0.002812; Train Acc 0.999643; Val Acc 0.924186\n",
            "Epoch 40; Loss 0.003898; Train Acc 0.998929; Val Acc 0.919381\n",
            "Epoch 41; Loss 0.004501; Train Acc 0.999286; Val Acc 0.923118\n",
            "Epoch 42; Loss 0.001364; Train Acc 0.998751; Val Acc 0.918847\n",
            "Epoch 43; Loss 0.004763; Train Acc 0.998394; Val Acc 0.911906\n",
            "Epoch 44; Loss 0.000597; Train Acc 0.999108; Val Acc 0.918847\n",
            "Epoch 45; Loss 0.001086; Train Acc 0.999643; Val Acc 0.918313\n",
            "Epoch 46; Loss 0.001979; Train Acc 0.999643; Val Acc 0.917245\n",
            "Epoch 47; Loss 0.000577; Train Acc 0.999643; Val Acc 0.919915\n",
            "Epoch 48; Loss 0.001019; Train Acc 0.999822; Val Acc 0.917779\n",
            "Epoch 49; Loss 0.000294; Train Acc 0.999822; Val Acc 0.920982\n",
            "Epoch 50; Loss 0.000262; Train Acc 0.999822; Val Acc 0.917245\n",
            "Epoch 51; Loss 0.000239; Train Acc 0.999822; Val Acc 0.917245\n",
            "Epoch 52; Loss 0.000219; Train Acc 0.999822; Val Acc 0.919381\n",
            "Epoch 53; Loss 0.000235; Train Acc 0.999822; Val Acc 0.918847\n",
            "Epoch 54; Loss 0.000368; Train Acc 0.999465; Val Acc 0.922584\n",
            "Epoch 55; Loss 0.000270; Train Acc 0.999822; Val Acc 0.921516\n",
            "Epoch 56; Loss 0.000203; Train Acc 0.999822; Val Acc 0.923652\n",
            "Epoch 57; Loss 0.000458; Train Acc 0.999643; Val Acc 0.924186\n",
            "Epoch 58; Loss 0.000856; Train Acc 0.999822; Val Acc 0.920982\n",
            "Epoch 59; Loss 0.000161; Train Acc 0.999822; Val Acc 0.922050\n",
            "Epoch 60; Loss 0.000120; Train Acc 1.000000; Val Acc 0.920982\n",
            "Epoch 61; Loss 0.000121; Train Acc 1.000000; Val Acc 0.920448\n",
            "Epoch 62; Loss 0.000080; Train Acc 1.000000; Val Acc 0.920448\n",
            "Epoch 63; Loss 0.000057; Train Acc 1.000000; Val Acc 0.920982\n",
            "Epoch 64; Loss 0.000047; Train Acc 1.000000; Val Acc 0.921516\n",
            "Epoch 65; Loss 0.000038; Train Acc 1.000000; Val Acc 0.920982\n",
            "Epoch 66; Loss 0.000116; Train Acc 1.000000; Val Acc 0.920982\n",
            "Epoch 67; Loss 0.000047; Train Acc 1.000000; Val Acc 0.920448\n",
            "Epoch 68; Loss 0.000039; Train Acc 1.000000; Val Acc 0.920448\n",
            "Epoch 69; Loss 0.000051; Train Acc 1.000000; Val Acc 0.920448\n",
            "Epoch 70; Loss 0.000031; Train Acc 1.000000; Val Acc 0.919915\n",
            "Epoch 71; Loss 0.000026; Train Acc 1.000000; Val Acc 0.920448\n",
            "Epoch 72; Loss 0.000027; Train Acc 1.000000; Val Acc 0.919915\n",
            "Epoch 73; Loss 0.000031; Train Acc 1.000000; Val Acc 0.920448\n",
            "Epoch 74; Loss 0.000052; Train Acc 1.000000; Val Acc 0.920982\n",
            "Epoch 75; Loss 0.000021; Train Acc 1.000000; Val Acc 0.919915\n",
            "Epoch 76; Loss 0.000033; Train Acc 1.000000; Val Acc 0.919915\n",
            "Epoch 77; Loss 0.000023; Train Acc 1.000000; Val Acc 0.919381\n",
            "Epoch 78; Loss 0.000024; Train Acc 1.000000; Val Acc 0.919381\n",
            "Epoch 79; Loss 0.000021; Train Acc 1.000000; Val Acc 0.919915\n",
            "Epoch 80; Loss 0.000043; Train Acc 1.000000; Val Acc 0.919915\n",
            "Epoch 81; Loss 0.000020; Train Acc 1.000000; Val Acc 0.920982\n",
            "Epoch 82; Loss 0.000021; Train Acc 1.000000; Val Acc 0.921516\n",
            "Epoch 83; Loss 0.000018; Train Acc 1.000000; Val Acc 0.920448\n",
            "Epoch 84; Loss 0.000022; Train Acc 1.000000; Val Acc 0.919915\n",
            "Epoch 85; Loss 0.000018; Train Acc 1.000000; Val Acc 0.920982\n",
            "Epoch 86; Loss 0.000017; Train Acc 1.000000; Val Acc 0.920448\n",
            "Epoch 87; Loss 0.000017; Train Acc 1.000000; Val Acc 0.919915\n",
            "Epoch 88; Loss 0.000012; Train Acc 1.000000; Val Acc 0.919915\n",
            "Epoch 89; Loss 0.000017; Train Acc 1.000000; Val Acc 0.919915\n",
            "Epoch 90; Loss 0.000014; Train Acc 1.000000; Val Acc 0.920448\n",
            "Epoch 91; Loss 0.000092; Train Acc 1.000000; Val Acc 0.920448\n",
            "Epoch 92; Loss 0.000019; Train Acc 1.000000; Val Acc 0.919381\n",
            "Epoch 93; Loss 0.000038; Train Acc 1.000000; Val Acc 0.923652\n",
            "Epoch 94; Loss 0.000044; Train Acc 1.000000; Val Acc 0.921516\n",
            "Epoch 95; Loss 0.000014; Train Acc 1.000000; Val Acc 0.919915\n",
            "Epoch 96; Loss 0.000024; Train Acc 1.000000; Val Acc 0.922584\n",
            "Epoch 97; Loss 0.000016; Train Acc 1.000000; Val Acc 0.922584\n",
            "Epoch 98; Loss 0.000013; Train Acc 1.000000; Val Acc 0.922584\n",
            "Epoch 99; Loss 0.000020; Train Acc 1.000000; Val Acc 0.920982\n",
            "Epoch 100; Loss 0.000016; Train Acc 1.000000; Val Acc 0.920982\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Final Model Predictions and Performance Metrics"
      ],
      "metadata": {
        "id": "ry08Q4j44tMQ"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Final Model and Accuracies"
      ],
      "metadata": {
        "id": "_cA8sMZj4uMb"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "From our hyperparameter tuning, we choose the model right above, which has the glove embeddings with size 300, and we chose the weights at checkpoint 57 (path is at checkpoint 56 because of 0 indexing), as it has the highest validation accuracy."
      ],
      "metadata": {
        "id": "rRljgBwQVP4o"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "glove = torchtext.vocab.GloVe(name=\"6B\", dim=300)\n",
        "model = Transformer(300).cuda()\n",
        "model.load_state_dict(torch.load('/content/gdrive/My Drive/checkpointing/ckpt-56.pk'))\n",
        "print('Train accuracy of the Transformer Model is: %f' % (get_accuracy(model, train_data)))\n",
        "print('Validation accuracy of the Transformer Model is: %f' % (get_accuracy(model, valid_data)))\n",
        "print('Test accuracy of the Transformer Model is: %f' % (get_accuracy(model, test_data)))"
      ],
      "metadata": {
        "id": "MqNNqs60WQ5f",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "f8b4ec8b-c4f4-4bac-bbf9-f61fccb73116"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "100%|█████████▉| 399999/400000 [00:37<00:00, 10597.81it/s]\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Train accuracy of the Transformer Model is: 0.999643\n",
            "Validation accuracy of the Transformer Model is: 0.924186\n",
            "Test accuracy of the Transformer Model is: 0.912879\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Finding one image that was classified correctly and one that was classified incorrectly in the test set"
      ],
      "metadata": {
        "id": "pbsMCT8ZdOD7"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test_loader = torch.utils.data.DataLoader(test_data, batch_size=1)\n",
        "model.eval()\n",
        "found_correct = False\n",
        "found_incorrect = False\n",
        "\n",
        "\n",
        "for post, label in test_loader:\n",
        "  post = post.cuda()\n",
        "  label = label.cuda()\n",
        "  output = model(post)\n",
        "  pred = output.argmax(1).item()\n",
        "  label = label.item()\n",
        "  if pred == label and not found_correct:\n",
        "    correct_post = post.tolist()[0] \n",
        "    found_correct = True\n",
        "  if pred != label and not found_incorrect:\n",
        "    incorrect_post = post.tolist()[0] \n",
        "    found_incorrect = True\n",
        "  if found_correct and found_incorrect:\n",
        "    break\n",
        "\n",
        "correct_post_words = [glove.itos[w] for w in list(filter(lambda a: a != 0, correct_post))]\n",
        "print(\"Correctly classified processed post in test set: \", correct_post_words)\n",
        "incorrect_post_words = [glove.itos[w] for w in list(filter(lambda a: a != 0, incorrect_post))]\n",
        "print(\"Incorrectly classified processed post in test set: \", incorrect_post_words)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qPpcC8kHd5IE",
        "outputId": "d8ac4558-24ee-4ad2-a784-900c6b7c7db2"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Correctly classified processed post in test set:  ['our', 'daily', 'update', 'is', 'published', '.', 'states', 'reported', 'tests', 'new', 'cases', 'and', '532', 'deaths', '.', 'current', 'hospitalizations', 'fell', 'below', '30k', 'for', 'first', 'time', 'since', 'june', '22', '.', '.']\n",
            "Incorrectly classified processed post in test set:  ['two', 'interesting', 'children', 'tend', 'to', 'weather', 'pretty', 'well', ';', 'they', 'also', 'get', 'a', 'ton', 'of', 'vitamin', 'd', '.', 'black', 'people', 'are', 'getting', 'slammed', 'by', ';', 'black', 'people', 'also', 'have', 'much', 'higher', 'instances', 'of', 'vitamin', 'd', 'deficiency', 'vs', 'in', 'general', '.']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Note from above's code the correctly classified pre-processed post is post #1 in the test set, which is: \"Our daily update is published. States reported 734k tests 39k new cases and 532 deaths. Current hospitalizations fell below 30k for the first time since June 22. https://t.co/wzSYMe0Sht\". This had label real, which our model correctly classified.\n",
        "\n",
        "\n",
        "\n",
        "The incorrectly classified pre-processed post from above's code is post #11 in the test set, which is: \"Two interesting correlations:\n",
        "\n",
        "1) Children tend to weather COVID-19 pretty well; they also get a ton of Vitamin D.\n",
        "\n",
        "2) Black people are getting slammed by COVID-19; black people also have much higher instances of Vitamin D deficiency (76% vs 40% in the general population).\". This had label fake, but our model misclassified it as real."
      ],
      "metadata": {
        "id": "1dIbuJEUtWqS"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## More Final Model Measures"
      ],
      "metadata": {
        "id": "p04Ft_-X4o9v"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "Getting the True Positive Rate (TPR), True Negative Rate (TNR), Positive Predictive Value (PPV), Negative Predictve Value (NPV) values on the test set."
      ],
      "metadata": {
        "id": "SGjXedQLcc2i"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def get_TP_TN_FP_FN(model, data):\n",
        "    '''\n",
        "    TP = true positive\n",
        "    TN = true negative\n",
        "    FP = false positive\n",
        "    FN = false negative\n",
        "    '''\n",
        "    loader = torch.utils.data.DataLoader(data, batch_size=256)\n",
        "    TP_tot, TN_tot, FP_tot, FN_tot = 0, 0, 0, 0\n",
        "    model.eval()\n",
        "    for posts, labels in loader:\n",
        "        posts = posts.cuda()\n",
        "        labels = labels.cuda()\n",
        "\n",
        "        output = model(posts)\n",
        "        pred = output.argmax(1)\n",
        "        \n",
        "        ones = torch.ones(pred.shape)\n",
        "        zeros = torch.zeros(pred.shape)\n",
        "        ones = ones.cuda()\n",
        "        zeros = zeros.cuda()\n",
        "        \n",
        "        correct = pred.eq(labels) \n",
        "        incorrect = (~pred).eq(labels)\n",
        "        positive = pred.eq(ones)\n",
        "        negative = pred.eq(zeros)\n",
        "        TP = correct[positive].sum().item()\n",
        "        TN = correct[negative].sum().item()\n",
        "        FP = positive.sum().item() - TP\n",
        "        FN = negative.sum().item() - TN\n",
        "        TP_tot += TP\n",
        "        TN_tot += TN\n",
        "        FP_tot += FP\n",
        "        FN_tot += FN \n",
        "    return TP_tot, TN_tot, FP_tot, FN_tot\n",
        "\n",
        "def get_TPR_FNR_TNR_FPR_PPV_FDR_NVP_FOR(TP, TN, FP, FN):\n",
        "    '''\n",
        "    TPR - true positive rate\n",
        "    FNR - false negative rate\n",
        "    TNR - true negative rate\n",
        "    FPR - false positive rate\n",
        "    PPV - positive predictive value\n",
        "    FDR - false discrovery rate\n",
        "    NPV - negative predictive value\n",
        "    FOR - false omission rate\n",
        "    '''\n",
        "    TPR = TP/(TP+FN)\n",
        "    FNR = FN/(TP+FN)\n",
        "    TNR = TN/(TN+FP)\n",
        "    FPR = FP/(TN+FP)\n",
        "    PPV = TP/(TP+FP)\n",
        "    FDR = FP/(TP+FP)\n",
        "    NPV = TN/(TN+FN)\n",
        "    FOR = FN/(TN+FN)\n",
        "    return TPR, FNR, TNR, FPR, PPV, FDR, NPV, FOR"
      ],
      "metadata": {
        "id": "wZ6PfWa0ywjG"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "TP, TN, FP, FN = get_TP_TN_FP_FN(model, test_data)\n",
        "TPR, FNR, TNR, FPR, PPV, FDR, NPV, FOR = get_TPR_FNR_TNR_FPR_PPV_FDR_NVP_FOR(TP, TN, FP, FN)\n",
        "print(TPR, TNR, PPV, NPV) "
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3lHACEP540CS",
        "outputId": "597d4b75-49ac-4d38-a224-3e540efb47be"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "0.9170353982300885 0.9088983050847458 0.9060109289617486 0.9196141479099679\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Viewing all the Incorrect Predictions to Look for Patterns"
      ],
      "metadata": {
        "id": "emEoJO_I82GF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test_loader = torch.utils.data.DataLoader(test_data, batch_size=1)\n",
        "model.eval()\n",
        "\n",
        "\n",
        "for post, label in test_loader:\n",
        "    post = post.cuda()\n",
        "    label = label.cuda()\n",
        "    output = model(post)\n",
        "    pred = output.argmax(1)\n",
        "\n",
        "    if pred != label:\n",
        "        incorrect_post = post.tolist()[0] \n",
        "        \n",
        "        print(\"pred: \", pred.item(), \"label: \", label.item(), [glove.itos[w] for w in list(filter(lambda a: a != 0, incorrect_post))])\n"
      ],
      "metadata": {
        "id": "6_3sLBh7-gje",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8d46a2ec-38ff-42aa-949c-4a591540f9fb"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "pred:  1 label:  0 ['two', 'interesting', 'children', 'tend', 'to', 'weather', 'pretty', 'well', ';', 'they', 'also', 'get', 'a', 'ton', 'of', 'vitamin', 'd', '.', 'black', 'people', 'are', 'getting', 'slammed', 'by', ';', 'black', 'people', 'also', 'have', 'much', 'higher', 'instances', 'of', 'vitamin', 'd', 'deficiency', 'vs', 'in', 'general', '.']\n",
            "pred:  1 label:  0 ['kids', 'reach', 'this', 'stage', 'of', 'lockdown', '.']\n",
            "pred:  1 label:  0 ['households', 'should', 'have', 'medical', 'kits', 'with', 'certain', 'items', 'and', 'equipment', 'to', 'treat', 'different', 'stages', 'of', '.']\n",
            "pred:  1 label:  0 ['health', 'minister', 'says', 'cases', 'with', 'no', 'confirmation', 'of', 'be', 'considered', 'on', 'death', 'toll', '.']\n",
            "pred:  0 label:  1 ['not', 'meaningless', 'but', 'also', 'before', 'this', 'epidemic', 'pneumonia', 'deaths', 'were', 'considered', 'a', 'very', 'unstable', 'number', '.', '.', '.', 'so', 'keep', 'that', 'in', 'mind', '.']\n",
            "pred:  0 label:  1 ['access', 'to', 'tools', 'accelerator', 'brings', 'together', 'governments', 'health', 'organizations', 'scientists', 'businesses', 'civil', 'society', ';', 'philanthropists', 'to', 'speed', 'up', 'an', 'end', 'to', 'pandemic', '.', '.', '.']\n",
            "pred:  0 label:  1 ['three', 'docs', 'have', 'been', 'getting', 'attention', 'for', 'their', 'letters', 'to', 'bmj', 'urging', 'more', 'attention', 'for', 'mild', 'yet', 'very', 'long', 'term', 'forms', 'of', '.', '.']\n",
            "pred:  0 label:  1 ['georgia', 'has', 'also', 'issued', 'a', 'stay', 'at', 'home', 'order', 'until', 'april', '13th', '.', '.']\n",
            "pred:  0 label:  1 ['nigeria', 'centre', 'for', 'disease', 'control', 'has', 'warned', 'nigerians', 'to', 'beware', 'of', 'scammers', 'posing', 'as', 'centre', 'to', 'defraud', 'unsuspecting', 'read', '.']\n",
            "pred:  0 label:  1 ['virtually', 'all', 'british', 'physicians', 'feel', 'concern', 'about', 'indirect', 'impact', 'of', 'coronavirus', 'on', 'their', 'patients', '.', '.', '.']\n",
            "pred:  0 label:  1 ['in', 'response', 'to', 'pandemic', 'pennsylvania', 'governor', 'tom', 'wolf', 'shut', 'down', 'nonessential', 'businesses', 'and', 'limited', 'gatherings', 'in', 'may', '.', 'a', 'federal', 'rules', 'that', 'these', 'restrictions', 'were', 'unconstitutional', 'violating', 'first', 'and', 'fourteenth', 'amendments', '.', '.']\n",
            "pred:  1 label:  0 ['smashes', 'record', 'for', 'most', 'new', 'cases', 'any', 'state', 'has', 'in', 'a', 'single', 'day', '.']\n",
            "pred:  1 label:  0 ['we', 'have', 'one', 'of', 'lowest', 'mortality', 'rates', 'in', 'world', 'from', '.']\n",
            "pred:  0 label:  1 ['un', 'secretary-general', 'antónio', 'guterres', 'has', 'told', 'sky', 'news', 'coronavirus', 'has', 'put', 'eradication', 'of', 'poverty', 'and', 'taken', 'some', 'aspects', 'of', 'progress', '25', 'years', 'back', '.', 'get', 'latest', 'news', '.', '.']\n",
            "pred:  0 label:  1 ['it', 'has', 'been', '77', 'days', 'since', 'last', 'case', 'of', 'was', 'acquired', 'locally', 'from', 'an', 'unknown', 'source', '.']\n",
            "pred:  0 label:  1 ['black', 'patients', 'more', 'likely', 'to', 'be', 'hospitalized', 'due', 'to', 'coronavirus', '.', '.', '.']\n",
            "pred:  1 label:  0 ['private', 'health', 'system', 'began', 'offering', 'antibody', 'tests', 'to', 'detect', '.', 'at', 'same', 'time', 'that', 'ministry', 'of', 'health', 'said', 'it', 'would', 'buy', 'antigen', 'tests', '.']\n",
            "pred:  0 label:  1 ['in', 'a', 'patient', 'with', 'a', 'skin', 'rash', 'is', 'a', 'useful', 'finding', 'to', 'help', 'diagnose', 'spanish', 'medical', 'experts', 'suggest', '.', '.', '.']\n",
            "pred:  0 label:  1 ['has', 'strengthened', 'negotiating', 'hand', '.', '.', '.']\n",
            "pred:  1 label:  0 ['rt', 'in', 'absence', 'of', 'clarity', 'and', 'a', 'rash', 'of', 'pandemic', 'has', 'created', 'a', 'breeding', 'ground', 'for']\n",
            "pred:  0 label:  1 ['also', 'intense', 'competition', 'for', 'samples', 'from', 'confirmed', 'cases', 'so', 'that', 'new', 'vendors', 'can', 'validate', 'their', 'tests', '.', 'we', 'need', 'a', 'nationwide', 'clearinghouse', 'for', 'samples', 'and', 'tests', '.']\n",
            "pred:  0 label:  1 ['is', 'a', 'celebrated', 'legal', 'scholar', '.', 'been', 'hosting', 'under', 'blacklight', 'at', 'to', 'look', 'at', 'intersectional', 'vulnerabilities', 'that', 'lays', 'bare', '.', '”', '.']\n",
            "pred:  0 label:  1 ['social', 'gatherings', 'of', 'more', 'than', 'six', 'people', 'are', 'now', 'illegal', 'in', 'england', 'as', 'of', 'comes', 'into', 'force', '.']\n",
            "pred:  1 label:  0 ['keep', 'your', 'newborn', 'more', 'than', '6', 'feet', 'away', 'from', 'you', 'as', 'much', 'as', 'possible', '.', 'discuss', 'with', 'your', 'healthcare', 'provider', 'about', 'using', 'a', 'physical', 'barrier', 'example', ',', 'placing', 'newborn', 'in', 'an', 'while', 'in', 'hospital', '.']\n",
            "pred:  0 label:  1 ['even', 'before', '14000', 'young', 'children', 'were', 'dying', 'every', 'day', '-', 'mostly', 'from', 'preventable', 'causes', '.', 'they', 'share', 'an', 'equal', 'right', 'to', 'survive', 'with', 'quality', 'health', 'care', '.', '.', '.']\n",
            "pred:  1 label:  0 ['this', 'year', 'election', 'night', 'could', 'be', 'chaotic', '.', 'in', 'addition', 'to', 'coping', 'with', 'an', 'unprecedented', 'number', 'of', 'mail', 'ballots', 'requires', 'election', 'officials', 'to', 'provide', 'clean', 'socially', 'distanced', 'in-person', 'voting', 'and', 'grapple', 'with', 'a', 'likely', 'shortage', 'of', 'election', 'workers', '.', '.']\n",
            "pred:  1 label:  0 ['luckily', 'sir', 'for', 'you', 'people', 'as', 'for', '95', '%', 'who', 'suffer', 'from', 'course', 'is', 'mild', '.', 'you', 'would', 'heal', 'even', 'without', 'taking', 'anything', '.', 'for', 'those', 'unlucky', '5', '%', 'who', 'actually', 'suffer', 'from', 'severe', 'and', 'have', 'ards', 'it', 'is', 'catastrophic', 'and', 'no', 'will', 'help', 'for', 'sure', '.']\n",
            "pred:  0 label:  1 ['percent', 'positive', 'has', 'been', 'a', 'critically', 'important', 'metric', 'but', 'flipping', 'fraction', 'can', 'help', 'us', 'communicate', 'more', 'persuasively', '.', 'and', 'explain', 'why', 'is', 'such', 'a', 'valuable', 'metric', '.', '.', '.']\n",
            "pred:  0 label:  1 ['donald', 'trump', 'has', 'claimed', 'he', 'seriousness', 'of', 'coronavirus', 'pandemic', '-', 'despite', 'admitting', 'earlier', 'this', 'year', 'he', 'had', 'to', 'always', 'play', 'it', '.']\n",
            "pred:  0 label:  1 ['since', 'coronavirus', 'there', 'has', 'been', 'a', 'drop', 'in', 'pediatric', 'ed', 'visits', '.', 'is', 'among', 'most', 'commonly', 'delayed', 'diagnoses', '.', '.']\n",
            "pred:  1 label:  0 ['200', 'new', 'cases', 'of', 'registered', 'in', 'lithuania', '.', 'people', 'are', 'with', 'at', 'vilnius', 'international', 'airport', 'and', 'taken', 'to', 'a', 'small', 'nearby', '.']\n",
            "pred:  1 label:  0 ['extraordinary', 'times', 'and', 'releasing', 'patients', 'to', 'curb', 'social', 'gatherings', '.']\n",
            "pred:  1 label:  0 ['48', 'positive', 'patients', 'from', 'found', 'in', 'sultanpuri', 'area', 'in', 'india', '.']\n",
            "pred:  1 label:  0 ['there', 'are', 'big', 'studies', 'and', 'findings', 'on', 'corresponding', 'relationships', 'between', 'and', 'vitamin', 'd', 'levels', 'in', 'patients', '.']\n",
            "pred:  1 label:  0 ['in', 'a', 'hungry', '&', 'jobless', 'youth', 'pour', 'petrol', 'on', 'himself', 'and', 'put', 'on', '.', '.', '.']\n",
            "pred:  0 label:  1 ['chairman', 'of', 'british', 'medical', 'association', 'dr', 'says', 'that', 'wearing', 'face', 'masks', 'is', 'explaining', 'that', 'they', 'prevent', '80', 'percent', 'of', 'virus', 'spreading', 'from', 'one', 'person', 'to', '.', '.', '.']\n",
            "pred:  1 label:  0 ['bank', 'of', 'america', 'said', 'funds', 'will', 'go', 'to', 'programs', 'that', 'assist', 'people', 'and', 'communities', 'of', 'color', 'affected', 'by', '.', 'areas', 'of', 'focus', 'will', 'include', 'health', 'job', 'training', 'skill', 'building', 'and', 'retraining', 'small', 'business', 'support', 'and', 'housing', 'assistance', '.', '.']\n",
            "pred:  0 label:  1 ['of', 'american', 'adults', 'believe', 'in', 'wearing', 'masks', '.', '.']\n",
            "pred:  0 label:  1 ['as', 'drivers', 'queue', 'for', 'hours', 'for', 'tests', 'before', 'being', 'turned', 'away', '.']\n",
            "pred:  1 label:  0 ['early', 'action', 'and', 'social', 'trust', 'are', 'among', 'reasons', 'for', 'low', 'numbers', 'of', 'coronavirus', 'cases', '.', '.']\n",
            "pred:  0 label:  1 ['pandemic', 'has', 'led', 'to', 'a', 'surge', 'in', 'opioid', 'overdose', 'deaths', 'in', 'one', 'hospital', 'a', 'new', 'study', 'confirms', '.', '.']\n",
            "pred:  1 label:  0 ['number', 'of', 'people', 'infected', 'with', 'rose', 'in', 'tokyo', 'because', 'there', 'was', 'no', 'longer', 'any', 'need', 'to', 'refrain', 'from', 'testing', 'after', 'postponement', 'of', 'olympics', '.']\n",
            "pred:  0 label:  1 ['h1n1', 'holds', 'lessons', 'for', 'an', 'efficient', 'delivery', 'of', 'vaccination', '.', '.']\n",
            "pred:  1 label:  0 ['florida', 'gov', '.', 'ron', 'desantis', 'said', 'today', 'kids', 'are', 'not', 'a', 'major', 'vector', 'of', 'spreading', '.', 'studies', 'show', 'that', 'high', 'schoolers', 'are', 'more', 'like', 'adults', 'while', 'young', 'children', 'present', 'a', 'lower', 'risk', '.', 'but', 'findings', 'overseas', 'may', 'not', 'apply', 'to', 'u', '.', 's', '.', 'states', 'with', 'rising', 'cases', '.', '.']\n",
            "pred:  1 label:  0 ['coronas', 'virus', 'must', 'be', 'stopped', 'confirm', 'irish', 'music', 'fans', '.']\n",
            "pred:  1 label:  0 ['live', 'talking', 'about', 'vaccine', 'with', '.', 'join', 'us', 'and', 'ask', 'some', 'questions', 'of', 'your', '.', '.']\n",
            "pred:  1 label:  0 ['all', 'india', 'lock', 'down', 'increased', 'till', '4th', 'may', '.', '.', '.']\n",
            "pred:  1 label:  0 ['this', 'is', 'to', 'tell', 'all', 'of', 'us', 'that', 'ph', 'for', 'coronary', 'virus', 'varies', 'from', '5', '.', '5', 'to', '8', '.', '5', '.']\n",
            "pred:  0 label:  1 ['scientists', 'without', 'trial', 'data', 'how', 'can', 'we', 'trust', 'vaccine', '?', '.', '.']\n",
            "pred:  0 label:  1 ['brazil', 'is', 'a', 'worrying', 'combination', 'of', 'pandemic', 'and', 'pandemonium', '.', '.']\n",
            "pred:  0 label:  1 ['change', 'avails', 'to', 'any', 'hospitalized', 'patients', 'not', 'just', 'severely', 'ill', '.', '.']\n",
            "pred:  1 label:  0 ['coronavirus', 'broadcast', 'strikes', 'dignified', 'tone', 'despite', 'zoom', 'background', '.']\n",
            "pred:  1 label:  0 ['convalescent', 'plasma', 'had', 'been', 'to', 'reduce', 'mortality', 'by', 'in', 'hospitalized', 'patients', '.']\n",
            "pred:  0 label:  1 ['coronavirus', 'is', 'accelerating', 'money', 'problems', '.', '.']\n",
            "pred:  1 label:  0 ['i', 'get', 'over', 'sitting', '15-20', 'feet', 'away', 'from', 'other', 'people', 'without', 'a', 'mask', 'is', 'a', 'problem', '.', 'then', 'again', 'cdc', 'will', 'tell', 'you', 'next', 'week', 'that', 'is', 'airborne', 'up', 'to', 'a', 'mile', '.']\n",
            "pred:  0 label:  1 ['what', 'do', 'you', 'need', 'to', 'know', 'about', 'if', 'you', 'work', 'with', 'people', 'with', 'in', '?', 'learn', '.', '.', '.']\n",
            "pred:  0 label:  1 ['first', 'documented', 'us', 'case', 'of', 'syndrome', 'associated', 'with', 'coronavirus', '.', '.', '.']\n",
            "pred:  1 label:  0 ['corona', 'time', '.', 'i', 'did', 'baba', 'yoga', 'for', '1', 'hour', 'every', 'day', '.', 'took', 'capsules', '.', 'deep', 'breathing', ';', 'sleeping', 'on', 'cheat', '.', 'these', 'saved', 'me', 'from', 'corona', '.']\n",
            "pred:  1 label:  0 ['a', '?', '?', '?', 'face', 'mask', 'exempt', 'card', '?', '?', 'allows', 'you', 'to', 'cite', 'ada', 'and', 'not', 'wear', 'a', 'mask', '.']\n",
            "pred:  0 label:  1 ['donald', 'trump', 'ignores', 'rules', 'with', 'and', 'indoor', 'rally', '.']\n",
            "pred:  0 label:  1 ['florida', 'tops', '100k', 'coronavirus', 'cases', '.', '.']\n",
            "pred:  1 label:  0 ['sir', 'i', 'was', 'infected', 'in', 'mid', 'august', 'of', '19', 'and', 'after', 'suffering', 'for', '8', 'days', 'was', 'prescribed', 'for', 'and', 'was', 'cured', '.', 'in', '36', 'hours', 'most', 'of', 'my', 'symptoms', 'were', 'gone', 'fully', 'recovered', 'in', '7', 'days']\n",
            "pred:  0 label:  1 ['earlier', 'in', 'month', 'director-general', 'of', 'health', 'signalled', 'a', 'move', 'to', 'more', 'strongly', 'focus', 'testing', 'at', 'our', 'borders', '.', 'that', 'approach', 'was', 'formally', 'announced', 'by', 'health', 'minister', 'this', 'week', '-']\n",
            "pred:  0 label:  1 ['in', 'pediatric', 'coronavirus', 'still', 'a', 'puzzle', '.', '.']\n",
            "pred:  0 label:  1 ['bolton', 'sees', 'another', 'sharp', 'rise', 'in', 'infections', '-', 'with', '192', 'new', 'cases', 'per', '100000', 'people', '.']\n",
            "pred:  1 label:  0 ['who', 'now', 'says', 'it', 'was', 'wrong', 'and', 'asymptomatic', 'spread', 'of', 'is', 'very', 'rare', '.']\n",
            "pred:  1 label:  0 ['of', 'payroll', 'data', 'presents', 'he', 'launched', 'earning', 'prospects', 'through', 'our', 'experts', 'get', 'more', 'companies', 'news', 'and', 'ease', 'of', 'baroda', 'verger', 'wockhardt', 'cipla', 'etc', 'are']\n",
            "pred:  1 label:  0 ['and', 'tweets', 'too', '.']\n",
            "pred:  1 label:  0 ['experts', 'at', 'nanavati', 'hospital', 'are', 'giving', 'four', 'common', 'treatments', 'to', 'all', 'vitamin', 'c', 'dry', 'ginger', 'turmeric', 'and', 'steam', '.']\n",
            "pred:  1 label:  0 ['wuhan', 'has', 'imposed', 'a', 'second', 'lockdown', 'after', 'a', 'resurgence', 'of', 'cases', '.']\n",
            "pred:  0 label:  1 ['a', 'karaoke', 'night', 'at', 'a', 'scottish', 'pub', 'has', 'been', 'linked', 'to', 'an', 'outbreak', 'of', 'coronavirus', 'cases', '.']\n",
            "pred:  1 label:  0 ['with', 'new', 'infections', 'in', 'last', '24', 'hours', 'tally', 'stands', 'at', '.', 'case', 'fatality', 'rate', 'now', 'stands', 'at', '2', '.', 'which', 'is', 'lowest', 'in', 'world', 'said', 'union', 'health', 'minister', 'dr', 'harsh', 'vardhan', '.', '.']\n",
            "pred:  1 label:  0 ['new', 'government', 'lockdown', 'advice', 'is', 'either', 'or', 'depending', 'on', 'who', 'you', 'voted', 'for', '.', '.']\n",
            "pred:  1 label:  0 ['human', 'coronavirus', 'infections', 'peak', 'in', '.', 'seem', 'to', 'peak', 'in', '.', 'elsevier', 'is', 'some', 'coronavirus', 'articles', 'again', '.', '.']\n",
            "pred:  1 label:  0 ['there', 'is', 'variation', 'in', 'mortality', 'and', 'infection', 'rate', 'based', 'upon', 'genome', 'of', 'virus', '.', 'host', 'immunity', 'is', 'also', 'playing', 'a', 'role', '.']\n",
            "pred:  1 label:  0 ['rt', 'remember', 'no', 'coronavirus', 'vaccine', 'is', 'ready', 'for', 'market', '.', 'here', 'are', 'vaccines', 'that', 'are', 'furthest', 'along', '.']\n",
            "pred:  1 label:  0 ['critics', 'say', 'it', 'is', 'wrong', 'to', 'use', 'of', 'pcr', 'tests', 'for', 'to', 'detect', 'since', 'in', 'theory', 'this', 'is', 'not', 'specific', 'to', '.', 'it', 'is', 'a', 'non-specific', 'test', 'in', 'which', 'genetic', 'material', 'from', 'any', 'organism', 'can', 'be', 'detected', 'and', 'can', 'potentially', 'misidentify', 'another', 'virus', 'as', '.']\n",
            "pred:  0 label:  1 ['rt', '500', 'bed', 'hospital', 'at', 'patna', 'will', 'be', 'inaugurated', 'today', 'and', '500', 'bed', 'hospital', 'at', 'muzaffarpur', 'will', 'be', 'inaugurated']\n",
            "pred:  0 label:  1 ['franchising', 'as', 'government', 'seeks', 'new', 'rail', 'future', '.']\n",
            "pred:  1 label:  0 ['nashville', 'woman', 'encouraged', 'to', 'learn', 'obesity', 'alone', 'does', 'not', 'appear', 'to', 'increase', 'risk', 'for', 'mortality', '.']\n",
            "pred:  0 label:  1 ['a', 'common', 'are', 'coronavirus', 'cases', 'going', 'up', 'because', 'testing', 'so', 'many', 'more', 'people', '?', 'certainly', 'not', 'in', 'florida', 'where', 'testing', 'slowed', 'down', 'while', 'new', 'cases', 'grew', 'over', 'last', 'week', '.', '.']\n",
            "pred:  1 label:  0 ['first', 'confirmed', 'case', 'traveled', 'to', 'taiwan', 'earlier', 'which', 'means', 'there', 'is', 'a', 'severe', 'community', 'transmission', 'in', 'taiwan', '.']\n",
            "pred:  0 label:  1 ['wait', 'this', 'is', 'not', 'time', 'series', 'right', '?', 'you', 'guys', 'are', 'giving', 'me', 'a', 'heart', 'attack', '.']\n",
            "pred:  1 label:  0 ['all', 'episodes', 'of', 'daily', 'to', 'be', 'available', 'to', 'on', 'bbc', 'iplayer', 'during', 'inevitable', 'second', 'peak', '.', '.']\n",
            "pred:  1 label:  0 ['roadmap', 'to', 'ease', 'restrictions', 'will', 'be', 'set', 'out', 'in', '5', 'phases', '.', 'these', 'phases', 'will', 'be', 'on', '3', 'week', 'review', 'process', '.']\n",
            "pred:  1 label:  0 ['restaurants', 'are', 'spitting', 'in', 'food', 'to', 'spread', 'coronavirus', '.']\n",
            "pred:  0 label:  1 ['you', 'can', 'i', 'get', 'from', 'food', 'including', 'delivery', 'or', 'restaurant', 'takeout', '?', 'learn', '.', '.', '.']\n",
            "pred:  1 label:  0 ['nhs', 'advises', 'to', 'wash', 'your', 'hands', 'like', 'you', 'just', 'picked', 'up', 'mark', 'dirty', '.', '.']\n",
            "pred:  0 label:  1 ['children', 'should', 'stay', 'at', 'school', 'until', 'there', 'is', 'a', 'positive', 'test', 'in', 'their', 'bubble', 'pm', 'says', '.']\n",
            "pred:  1 label:  0 ['handy', 'tips', 'to', 'stay', 'productive', 'while', 'in', '.']\n",
            "pred:  1 label:  0 ['community', 'steam', 'inhalation', 'or', 'smokers', '.']\n",
            "pred:  0 label:  1 ['this', 'morning', 'who', 'director-general', 'dr', 'has', 'emphasised', 'that', 'while', 'a', 'number', 'of', 'vaccines', 'are', 'in', 'clinical', 'trials', 'there', 'is', 'currently', 'no', 'silver', 'bullet', 'for', '.']\n",
            "pred:  0 label:  1 ['coronavirus', 'hospital', 'admissions', 'in', 'england', 'reach', 'highest', 'level', 'since', 'july', '.', 'follow', 'latest', 'updates', '.']\n",
            "pred:  1 label:  0 ['decided', 'to', 'opt', 'out', 'of', 'kcl', 'symptoms', 'tracker', 'study', '.', 'it', 'made', 'me', 'resent', 'that', 'a', 'majority', 'of', 'other', 'health', 'issues', 'have', 'been', 'sidelined', 'thanks', 'to', 'this', 'virus', '.', 'i', 'understand', 'that', 'virus', 'is', 'dangerous', '.', 'but', 'nhs', 'has', 'also', 'shut', 'down', 'to', 'prevent', 'spread', '.']\n",
            "pred:  0 label:  1 ['care', 'homes', 'boss', 'says', 'government', 'has', 'been', 'and', 'over', 'fears', '.']\n",
            "pred:  0 label:  1 ['rt', 'we', 'have', 'asked', 'world', 'health', 'organization', 'for', '1', 'million', 'coronavirus', 'test', 'kits', 'says', 'icmr', 'dr', 'balram', 'bhargava']\n",
            "pred:  1 label:  0 ['cases', 'up', 'only', 'because', 'of', 'our', 'big', 'number', 'testing', '.', 'mortality', 'rate', 'way']\n",
            "pred:  1 label:  0 ['rt', 'wisconsin', 'is', 'trending', 'poorly', 'on', 'cases', 'but', 'gov', '.', 'tony', 'stat', 'quite', 'right', '.']\n",
            "pred:  1 label:  0 ['home', 'gyms', 'struggle', 'to', 'retain', 'members', 'after', 'trial', 'period', 'ends', '.', '.']\n",
            "pred:  0 label:  1 ['why', 'did', 'go', 'away', 'while', 'keeps', 'on', 'spreading', '?', '.']\n",
            "pred:  1 label:  0 ['prisons', 'to', 'release', 'offenders', 'early', 'to', 'prevent', 'coronavirus', 'spread', 'in', 'us', 'penal', 'system', '.']\n",
            "pred:  1 label:  0 ['aiims', 'has', 'released', 'a', 'list', 'of', 'respiratory', 'symptoms', 'of', 'and', 'similar', 'diseases', '.']\n",
            "pred:  0 label:  1 ['as', 'of', '23', 'july', '2020', '9:54', 'pm', 'there', 'are', 'people', 'under', 'in', '.']\n",
            "pred:  1 label:  0 ['new', 'discovery', 'changes', 'things', '.']\n",
            "pred:  0 label:  1 ['while', 'farcical', 'disinformation', 'about', 'coronavirus', 'is', 'easily', 'dismissed', 'by', 'medical', 'professions', 'your', 'patients', 'may', 'get', 'sucked', 'in', '.', '.', '.']\n",
            "pred:  0 label:  1 ['cto', 'is', 'an', 'important', 'new', 'md', 'title', 'created', 'to', 'address', 'coronavirus', '.', '.']\n",
            "pred:  1 label:  0 ['hydrogels', 'are', 'part', 'of', 'mrna', 'delivery', 'system', '.', 'hydrogels', 'will', 'connect', 'you', 'to', 'internet', '.']\n",
            "pred:  1 label:  0 ['while', 'tests', 'may', 'sometimes', 'produce', 'false-positive', 'results', 'rare', '.', 'experts', 'are', 'more', 'concerned', 'about', '.', '.']\n",
            "pred:  0 label:  1 ['rt', 'a', 'leading', 'coronavirus', 'model', 'has', 'upped', 'its', 'predicted', 'death', 'toll', 'again', 'this', 'time', 'projecting', 'americans', 'will', 'lose', 'their']\n",
            "pred:  1 label:  0 ['rt', 'just', 'like', 'we', 'wash', 'our', 'hands', 'to', 'fight', 'we', 'can', 'scrub', 'our', 'chats', 'to', 'contain', '.', 'here', 'are', 'some']\n",
            "pred:  1 label:  0 ['un', 'secretary-general', 'announcement', 'that', 'vaccine', 'trials', 'will', 'begin', 'in', 'africa', '.']\n",
            "pred:  0 label:  1 ['at', 'this', 'rate', 'india', 'is', 'slated', 'to', 'overtake', 'in', 'testing', '.']\n",
            "pred:  1 label:  0 ['people', 'are', 'starting', 'to', 'enter', 'ers', 'with', 'fungal', 'lung', 'infections', 'from', 'wearing', 'take', 'breaks', 'from', 'your', 'masks', '.']\n",
            "pred:  1 label:  0 ['models', 'projecting', 'deaths', 'talking', 'about', 'without', 'mitigation', '.', '”']\n",
            "pred:  0 label:  1 ['extremely', 'unlikely', 'given', 'positive', 'test', 'rate', 'of', 'u', '.', 's', '.', 'and', 'known', 'testing', 'criteria', '.', 'see', 'among', 'other', '.']\n",
            "pred:  0 label:  1 ['as', 'per', 'after', 'completion', 'of', 'period', 'it', 'is', 'advised', 'to', 'contact', 'field', 'officer', 'for', 'issuance', 'of', 'a', 'fitness', 'certificate', '.', 'there', 'is', 'no', 'need', 'for', 'testing', 'after', 'home', 'period', 'is', '.', '.']\n",
            "pred:  0 label:  1 ['slovenia', 'and', 'guadeloupe', 'added', 'to', 'quarantine', 'list', '-', 'but', 'thailand', 'and', 'singapore', 'removed', '.']\n",
            "pred:  0 label:  1 ['people', 'are', 'drinking', 'sanitizer', 'to', 'get', 'an', 'alcohol', 'high', 'a', 'dangerous', 'trend', '.', '.']\n",
            "pred:  0 label:  1 ['our', 'national', 'contact', 'tracing', 'system', 'has', 'recorded', '85', 'close', 'contacts', 'identified', 'from', 'three', 'gym', 'classes', 'at', 'les', 'mills', 'takapuna', '–', 'they', 'have', 'all', 'been', 'contacted', 'and', 'are', '.']\n",
            "pred:  1 label:  0 ['20', 'countries', 'with', 'most', 'cases', 'together', 'home', 'to', 'as', 'many', 'people', 'as', 'india', 'have', 'conducted', 'tests', 'that', 'india', 'has', '.', 'our', '.']\n",
            "pred:  0 label:  1 ['rt', 'cover', 'your', 'spreads', 'through', 'respiratory', 'droplets', 'when', 'an', 'infected', 'person', 'coughs', 'sneezes', 'or', 'talks', '.']\n",
            "pred:  1 label:  0 ['coronavirus', 'pandemic', 'may', 'be', 'cut', 'short', 'due', 'to', 'short', 'attention', 'span', '.']\n",
            "pred:  0 label:  1 ['all', 'infants', 'born', 'to', 'a', 'cohort', 'of', '31', 'mothers', 'tested', 'negative', 'for', 'virus', 'during', 'height', 'of', 'new', 'york', 'surge', '.', '.']\n",
            "pred:  0 label:  1 ['affects', 'people', 'in', 'all', 'communities', '.', 'award-winning', 'actor', 'and', 'tribal', 'elder', 'wes', 'studi', 'offers', 'guidance', 'to', 'tribal', 'communities', 'to', 'help', 'slow', 'spread', 'of', 'coronavirus', 'or', 'in', 'navajo', 'language', '.', '.', '.']\n",
            "pred:  0 label:  1 ['and', 'today', 'it', 'is', 'also', 'looking', 'like', 'it', 'will', 'be', 'over', '.', 'i', 'think', 'this', 'is', 'really', 'a', 'winning', 'argument', 'for', 'people', 'making', 'it', 'but', 'intent', 'seem', 'to', 'actually', 'be', 'accuracy', 'but', 'sowing', 'confusion', '.']\n",
            "pred:  0 label:  1 ['also', 'a', 'debate', 'over', 'which', 'to', 'target', '.', 'seen', 'total', 'antibodies', 'igm', 'igg', 'and', 'iga', '.', 'each', 'has', 'a', 'different', 'time', 'course', 'and', 'optimal', 'type', 'of', 'sample', '.', 'which', 'is', 'best', 'for', 'which', 'purpose', '?']\n",
            "pred:  0 label:  1 ['up', 'bright', 'and', 'early', 'celebrating', 'healthcare', 'workers', 'everywhere', '.', 'we', 'be', 'able', 'to', 'stop', 'thanking', 'you', 'for', 'being', 'on', 'frontline', 'fighting', '.', 'thank', 'you', 'again', 'and', 'again', 'and', 'again', '.']\n",
            "pred:  0 label:  1 ['fda', 'official', 'says', 'if', 'a', 'vaccine', 'is', 'approved', 'before', 'ready', '--', 'outta', 'there', '.', '.', '.']\n",
            "pred:  1 label:  0 ['supreme', 'leader', 'ayatollah', 'ali', 'khamenei', 'said', 'us', 'created', 'special', 'of', 'virus', 'that', 'is', 'built', 'for', 'iran', 'using', 'genetic', 'data', 'of', 'iranians', 'which', 'they', 'have', 'obtained', 'through', 'different']\n",
            "pred:  1 label:  0 ['northern', 'ireland', 'was', 'testing', 'for', 'at', 'a', 'rate', '10', 'times', 'that', 'of', 'scotland', 'reported', 'on', '9', 'may', '2020', '.']\n",
            "pred:  1 label:  0 ['not', 'a', '4', 'to', 'carry', 'a', 'firearm', 'while', 'wearing', 'a', 'face', 'mask', 'to', 'prevent', 'spread', 'of', '.']\n",
            "pred:  0 label:  1 ['coronavirus', 'second', 'pm', 'considers', 'new', 'measures', 'as', 'london', 'mayor', 'insists', 'should', 'not', '.']\n",
            "pred:  0 label:  1 ['uk', 'alert', 'level', 'should', 'be', 'upgraded', 'chief', 'medical', 'officers', 'say', '.']\n",
            "pred:  0 label:  1 ['we', 'are', 'now', 'seeing', 'a', 'second', 'wave', 'coming', 'in', '.', 'prime', 'minister', 'boris', 'johnson', 'says', 'confirms', 'uk', 'is', 'seeing', 'a', 'second', 'wave', 'adding', 'that', 'it', 'was', 'always', '.', \"'\", 'more', '.', '.']\n",
            "pred:  1 label:  0 ['cdc', 'finally', 'confirmed', 'virus', 'is', 'airborne', '.', 'not', 'enough', 'to', 'maintain', 'social', 'distancing', 'and', 'wash', 'your', 'hands', '.', 'you', 'must', 'wear', 'a', 'mask', 'and', 'avoid', 'indoor', 'public', 'spaces', 'as', 'much', 'as', 'possible', '.']\n",
            "pred:  0 label:  1 ['conjunctivitis', 'may', 'sometimes', 'be', 'only', 'coronavirus', 'symptom', '.', '.', '.']\n",
            "pred:  0 label:  1 ['chinese', 'scientists', 'are', 'racing', 'to', 'start', 'human', 'trials', 'for', 'a', 'vaccine', 'by', 'august', '.', '.', '.']\n",
            "pred:  0 label:  1 ['labour', 'leader', 'sir', 'keir', 'starmer', 'is', 'after', 'a', 'member', 'of', 'his', 'household', 'displayed', 'possible', 'coronavirus', 'symptoms', '.', 'get', 'latest', 'on', 'this', 'story', '.', '.']\n",
            "pred:  0 label:  1 ['if', 'been', 'in', 'close', 'contact', 'with', 'a', 'person', 'who', 'has', 'a', 'public', 'health', 'worker', 'may', 'call', 'to', 'let', 'you', 'know', 'been', 'exposed', '.', 'answer', 'call', 'to', '.', 'learn', 'what', 'happens', 'during', 'contact', '.', '.', '.']\n",
            "pred:  1 label:  0 ['now', 'in', '400', 'mg', '.', '.']\n",
            "pred:  0 label:  1 ['labour', 'leader', 'sir', 'keir', 'starmer', 'after', 'household', 'member', 'showed', 'symptoms', '.']\n",
            "pred:  1 label:  0 ['if', 'you', 'have', 'symptoms', 'of', 'take', 'azithromycin', 'ivermectin', 'and', 'acetylcysteine', '.']\n",
            "pred:  1 label:  0 ['tinder', 'adds', 'new', 'option', 'for', 'users', '.']\n",
            "pred:  0 label:  1 ['.', '.', '.']\n",
            "pred:  1 label:  0 ['there', 'were', 'pcr', 'tests', 'in', 'tokyo', 'on', 'may', '19th', '.']\n",
            "pred:  1 label:  0 ['virus', 'was', 'engineered', 'and', 'optimised', 'for', 'human', 'transmission', 'by', 'humans', '.', 'their', 'zoonotic', 'cover', 'story', 'has', 'been', 'trashed', '.', 'having', 'created', 'a', 'pandemic', 'they', 'want', 'to', 'put', 'blame', 'for', 'it', 'on', 'someone', 'else', '.', 'this', 'explanation', 'is', 'not', 'entirely', 'trustworthy', '.', 'follow', 'money', '.']\n",
            "pred:  1 label:  0 ['two', 'members', 'of', 'test', 'positive', 'ahead', 'of', 'not', 'sure', 'how', 'rest', 'of', 'team', 'can', 'be', 'on', 'field', 'in', 'these']\n",
            "pred:  1 label:  0 ['a', 'study', 'from', 'cdc', 'and', 'who', 'proves', 'face', 'masks', 'do', 'not', 'prevent', 'spread', 'of', 'a', 'virus', '.']\n",
            "pred:  0 label:  1 ['first', 'case', 'is', 'a', 'woman', 'in', 'her', '30s', 'who', 'arrived', 'from', 'dubai', 'on', '28', 'august', 'second', 'is', 'a', 'child', 'who', 'arrived', 'from', 'uzbekistan', 'via', 'dubai', 'on', '28', 'august', '.', 'both', 'cases', 'are', 'now', 'being', 'transferred', 'to', 'auckland', 'quarantine', 'facility', 'with', 'other', 'members', 'of', 'their', 'bubble', '.']\n",
            "pred:  0 label:  1 ['estimated', 'size', 'of', 'unsustainable', 'corporate', 'debt', 'mountain', 'created', 'by', 'pandemic', 'is', 'to', 'be', 'slashed', 'to', 'by', 'influential', 'city', 'figures', '.']\n",
            "pred:  0 label:  1 ['offering', 'essential', 'cancer', 'care', 'in', 'new', 'normal', 'of', 'coronavirus', '.', '.', '.']\n",
            "pred:  0 label:  1 ['given', 'increased', 'spread', 'of', 'novel', 'coronavirus', 'across', 'world', 'cdc', 'has', 'updated', 'definition', 'of', 'a', 'person', 'under', 'investigation', 'for', '.', '.', '.']\n",
            "pred:  0 label:  1 ['hospital', 'acquired', 'coronavirus', 'can', 'reach', 'staff', 'as', 'well', 'as', 'patients', '.', '.', '.']\n",
            "pred:  0 label:  1 ['nejm', 'letter', 'says', 'saliva', 'test', 'as', 'good', 'or', 'better', 'than', 'nasal', 'swab', 'for', 'test', '.', '.']\n",
            "pred:  0 label:  1 ['arterial', 'and', 'venous', 'thrombosis', 'issues', 'with', 'coronavirus', 'and', 'role', 'of', 'anticoagulation', '.', '.', '.']\n",
            "pred:  0 label:  1 ['yeah', 'i', 'mean', '.', 'it', 'makes', 'sense', 'though', 'i', 'am', 'hearing', 'much', 'lower', 'numbers', 'from', 'a', 'couple', 'of', 'laboratories', 'and', 'one', 'private', 'company', '.', 'so', 'possible', 'if', 'you', 'have', 'right', 'setup', '.']\n",
            "pred:  0 label:  1 ['had', 'some', 'illinois', 'reporters', 'wonder', 'aloud', 'about', 'numbers', '.', 'we', 'have', 'not', 'seen', 'reason', 'to', 'doubt', 'them', 'than', 'for', 'any', 'other', 'state', 'but', 'just', '.']\n",
            "pred:  0 label:  1 ['positive', 'days', 'are', 'usually', 'when', 'a', 'state', 'is', 'reporting', 'negatives', 'out', 'of', 'cadence', 'with', 'positives', '.', 'seen', 'a', 'lot', 'of', 'states', 'do', 'this', 'recently', '.']\n",
            "pred:  1 label:  0 ['new', 'york', 'hospitals', 'reporting', 'fungal', 'lung', 'infections', 'from', 'face', 'masks', 'you', 'should', 'wear', 'to', 'slow', 'spread', 'of', '.', '.']\n",
            "pred:  1 label:  0 ['a', 'video', 'post', 'claims', 'rt-pcr', 'tests', 'used', 'to', 'detect', 'an', 'appropriate', 'testing', 'method', '.']\n",
            "pred:  0 label:  1 ['based', 'on', 'what', 'is', 'known', 'about', 'how', 'novel', 'coronavirus', 'behaves', 'cdc', 'and', 'expect', 'to', 'see', 'infection', 'in', 'people', 'who', 'have', 'had', 'contact', 'with', 'patient', 'especially', 'those', 'who', 'have', 'had', 'close', 'prolonged', 'contact', '.']\n",
            "pred:  1 label:  0 ['update', '.']\n",
            "pred:  0 label:  1 ['a', 'possibility', 'for', 'patients', 'to', 'wean', 'off', 'ventilator', 'or', 'avoid', 'it', 'altogether', '.', '.']\n",
            "pred:  0 label:  1 ['jedward', 'call', 'on', 'celebrities', 'and', 'influencers', 'to', 'speak', 'out', 'about', 'wearing', 'masks', '.']\n",
            "pred:  0 label:  1 ['a', 'major', 'concern', 'is', 'that', 'specificity', 'is', 'not', 'tested', 'using', 'samples', 'known', 'to', 'have', 'antibodies', 'against', 'antigens', 'from', 'one', 'or', 'more', 'of', '4', 'coronaviruses', 'that', 'cause', 'common', 'cold', '.']\n",
            "pred:  0 label:  1 ['rt', '?', 'one-stop', 'shop', 'for', 'resources', 'has', 'a', 'section', 'for', 'communicating', 'with', 'people', 'ages', '15-21', '.', 'find', 'those']\n",
            "pred:  0 label:  1 ['more', 'than', 'half', 'of', 'pregnant', 'women', 'recently', 'admitted', 'to', 'a', 'uk', 'hospital', 'with', 'infection', 'were', 'from', 'black', 'or', 'other', 'ethnic', 'minority', 'groups', 'bmj', 'study', 'finds', '.', '.']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Viewing all the Correct Predictions to Look for Patterns"
      ],
      "metadata": {
        "id": "jTCm0h3u8qoP"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "test_loader = torch.utils.data.DataLoader(test_data, batch_size=1)\n",
        "model.eval()\n",
        "\n",
        "\n",
        "for post, label in test_loader:\n",
        "    post = post.cuda()\n",
        "    label = label.cuda()\n",
        "    output = model(post)\n",
        "    pred = output.argmax(1)\n",
        "\n",
        "    if pred == label:\n",
        "        correct_post = post.tolist()[0]\n",
        "    \n",
        "        print(\"pred: \", pred.item(), \"label: \", label.item(), [glove.itos[w] for w in list(filter(lambda a: a != 0, correct_post))])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "o9dY4MCvWYvN",
        "outputId": "cb7072ae-401d-459f-9dac-47b2b9a72f0e"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'states', 'reported', 'tests', 'new', 'cases', 'and', '532', 'deaths', '.', 'current', 'hospitalizations', 'fell', 'below', '30k', 'for', 'first', 'time', 'since', 'june', '22', '.', '.']\n",
            "pred:  0 label:  0 ['alfalfa', 'is', 'only', 'cure', 'for', '.']\n",
            "pred:  0 label:  0 ['president', 'trump', 'asked', 'what', 'he', 'would', 'do', 'if', 'he', 'were', 'to', 'catch', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['states', 'reported', '630', 'deaths', '.', 'we', 'are', 'still', 'seeing', 'a', 'solid', 'national', 'decline', '.', 'death', 'reporting', 'lags', 'approximately', '28', 'days', 'from', 'symptom', 'onset', 'according', 'to', 'cdc', 'models', 'that', 'consider', 'lags', 'in', 'symptoms', 'time', 'in', 'hospital', 'and', 'death', 'reporting', 'process', '.', '.']\n",
            "pred:  1 label:  1 ['this', 'is', 'sixth', 'time', 'a', 'global', 'health', 'emergency', 'has', 'been', 'declared', 'under', 'international', 'health', 'regulations', 'but', 'it', 'is', 'easily', 'most', '.']\n",
            "pred:  1 label:  1 ['low', 'was', 'an', 'independent', 'predictor', 'of', 'worse', 'prognosis', 'in', 'patients', 'with', '.', '.', '.']\n",
            "pred:  1 label:  1 ['a', 'common', 'why', 'are', 'cumulative', 'outcome', 'numbers', 'smaller', 'than', 'current', 'outcome', 'numbers', '?', 'most', 'states', 'report', 'current', 'but', 'a', 'few', 'states', 'report', 'cumulative', '.', 'they', 'are', 'apples', 'and', 'oranges', 'and', 'we', 'feel', 'comfortable', 'filling', 'in', 'state', 'cumulative', 'boxes', 'with', 'current', '#s', '.']\n",
            "pred:  1 label:  1 ['government', 'should', 'consider', 'bringing', 'in', 'any', 'new', 'national', 'lockdown', 'rules', 'over', 'christmas', 'rather', 'than', 'now', 'says', 'an', 'oxford', 'university', 'professor', '.']\n",
            "pred:  0 label:  0 ['a', 'photo', 'shows', 'a', '19-year-old', 'vaccine', 'for', 'canine', 'coronavirus', 'that', 'could', 'be', 'used', 'to', 'prevent', 'new', 'coronavirus', 'causing', '.']\n",
            "pred:  1 label:  1 ['assistant', 'undersecretary', 'for', 'public', 'health', 'affairs', 'dr', 'highlighted', 'importance', 'of', 'risk', 'communication', ';', 'public', 'education', 'in', 'fighting', '.', 'country', 'increased', 'testing', 'capacity', 'enabling', 'people', 'to', 'be', 'tested', '.']\n",
            "pred:  0 label:  0 ['says', 'coronavirus', 'aid', ',', 'relief', ',', 'and', 'economic', 'security', 'act', 'gives', 'members', 'of', 'congress', 'a', 'pay', 'increase', '.']\n",
            "pred:  0 label:  0 ['an', 'image', 'of', 'a', 'man', 'carrying', 'his', 'old', 'mother', 'on', 'his', 'back', 'shows', 'migrants', 'travelling', 'to', 'their', 'homes', 'in', 'india', 'amidst', 'lockdown', '.']\n",
            "pred:  1 label:  1 ['3/10', 'about', 'of', 'population', 'be', 'infected', 'and', 'have', '.', 'of', 'people', 'are', 'still', 'vulnerable', 'to', 'disease', 'says', 'sir', 'patrick', 'vallance', '.', '.', '.']\n",
            "pred:  0 label:  0 ['a', '2009', 'tweet', 'from', 'donald', 'trump', 'said', 'that', 'he', 'would', 'never', 'let', 'thousands', 'of', 'americans', 'die', 'from', 'a', 'pandemic', 'while', 'in', 'office', '.']\n",
            "pred:  1 label:  1 ['icus', 'are', 'full', 'forcing', 'patients', 'to', 'wait', 'and', 'admit', 'to', 'emergency', 'room', '.', '.']\n",
            "pred:  1 label:  1 ['new', 'york', 'continues', 'to', 'have', 'highest', 'positive', 'tests', 'per', 'capita', 'an', 'indication', 'of', 'both', 'intensity', 'of', 'testing', 'there', 'and', 'severity', 'of', 'outbreak', '.', 'top', 'new', 'york', 'washington', 'new', 'jersey', 'louisiana', 'dc', 'michigan', 'illinois', 'vermont', 'colorado', 'rhode', 'island']\n",
            "pred:  0 label:  0 ['govt', 'has', 'added', 'disease', 'in', 'all', 'existing', 'insurances', 'as', 'a', 'special', 'case', '.']\n",
            "pred:  1 label:  1 ['she', 'has', 'been', 'in', 'managed', 'isolation', 'at', 'distinction', 'hotel', 'in', 'hamilton', 'and', 'tested', 'positive', 'around', 'day', '3', 'of', 'her', 'stay', '.', 'we', 'have', '13', 'new', 'confirmed', 'cases', 'in', 'auckland', '.', 'they', 'are', 'all', 'linked', 'to', '4', 'people', 'who', 'we', 'reported', 'as', 'cases', 'yesterday', 'and', 'day', 'before', '.']\n",
            "pred:  0 label:  0 ['gates', 'foundation', 'patent', 'for', 'this', 'coronavirus', '.', '\"']\n",
            "pred:  0 label:  0 ['nashville', 'man', 'thinks', 'world', 'is', 'upside-down', 'but', 'respects', 'that', 'world', 'care', 'what', 'he', 'thinks', '.']\n",
            "pred:  1 label:  1 ['there', 'are', 'also', 'clear', 'constraints', 'as', 'testing', 'criteria', 'remain', 'quite', 'strict', '.', 'we', 'are', 'still', 'getting', 'reports', 'from', 'all', 'over', 'of', 'people', 'who', 'have', 'been', 'able', 'unable', 'to', 'get', 'tested', 'despite', 'having', 'a', 'good', 'reason', 'to', 'do', 'so', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'found', 'in', 'or', 'chewing', 'tobacco', 'samples', 'sold', 'in', 'indian', 'state', 'of', 'maharashtra', 'as', 'raw', 'material', 'sourced', 'from', 'china', '.']\n",
            "pred:  1 label:  1 ['clinical', 'senior', 'lecturer', 'at', 'college', 'london', 'dr', 'claire', 'steves', 'says', 'symptoms', 'of', 'that', 'are', 'not', 'on', 'core', 'list', 'are', 'headaches', 'and', 'and', 'discusses', 'what', 'symptoms', 'warrant', 'a', 'test', 'for', '.', '.']\n",
            "pred:  1 label:  1 ['kia', 'ora', 'elanor', 'there', 'are', 'currently', '21', 'active', 'cases', 'in', 'new', 'zealand', 'they', 'were', 'all', 'caught', 'at', 'border', 'and', 'we', 'have', 'no', 'evidence', 'of', 'community', 'transmission', '.']\n",
            "pred:  1 label:  1 ['update', 'from', 'there', 'are', '3', 'new', 'cases', 'of', 'to', 'report', 'in', 'managed', 'isolation', 'and', 'quarantine', 'facilities', 'in', 'nz', '.', 'there', 'continue', 'to', 'be', 'no', 'new', 'cases', 'in', 'community', '.', 'our', 'total', 'number', 'of', 'active', 'cases', 'is', '21', '-', 'all', 'of', 'which', 'remain', 'in', 'quarantine', 'facilities', '.']\n",
            "pred:  0 label:  0 ['new', 'coronavirus', 'does', 'not', 'settle', 'in', 'air', 'but', 'is', 'grounded', 'so', 'it', 'is', 'not', 'transmitted', 'by', 'air', '.']\n",
            "pred:  1 label:  1 ['if', 'you', 'have', 'signs', 'of', 'a', 'heart', 'attack', 'or', 'stroke', 'while', 'staying', 'at', 'home', 'during', 'call', '9-1-1', 'right', 'away', '.', 'emergency', 'responders', 'and', 'emergency', 'departments', 'have', 'plans', 'in', 'place', 'to', 'help', 'protect', 'you', 'from', '.', 'learn', '.', '.', '.']\n",
            "pred:  0 label:  0 ['says', 'cdc', 'now', 'says', 'that', 'coronavirus', 'can', 'survive', 'on', 'surfaces', 'for', 'up', 'to', '17', 'days', '.']\n",
            "pred:  0 label:  0 ['chinese', 'government', 'announced', 'that', 'is', 'a', 'preventative', 'food', 'for', 'novel', 'coronavirus', '.', '\"']\n",
            "pred:  0 label:  0 ['flood', 'hit', 'areas', 'pretend', 'to', 'have', 'coronavirus', 'in', 'bid', 'to', 'get', 'attention', '.']\n",
            "pred:  0 label:  0 ['as', 'business', 'continue', 'to', 'adapt', 'their', 'services', 'in', 'wake', 'of', ',', 'swa', 'hand', 'portal', 'is', 'available', 'for', 'sourcing', 'and', 'ethanol', 'for', 'producing', '.', 'complete', 'your', 'details', 'and', 'find', 'suppliers', '.', '.']\n",
            "pred:  1 label:  1 ['free', 'watch', 'new', 'on', 'acute', 'flaccid', 'myelitis', '.', 'listen', 'to', 'our', 'new', 'podcast', 'as', 'dr', '.', 'john', 'iskander', 'and', 'dr', '.', 'janell', 'routh', 'discuss', 'afm', 'its', 'symptoms', 'and', 'treatment', 'options', '.', '.', '.']\n",
            "pred:  0 label:  0 ['.', 'virus', 'cure', '.', 'how', 'fast', 'does', 'this', 'pandemic', 'end', 'if', 'you', 'could', 'start', 'feeling', 'better', 'in', 'as', 'soon', 'as', '5', 'hours', '?']\n",
            "pred:  0 label:  0 ['subhan', 'after', 'corona', 'virus', 'china', 'govt', 'lifted', 'ban', 'on', 'holy', 'quran', '&', 'allowed', 'chinese', 'muslims', 'to', 'read', 'their', 'sacred', 'which', 'of', 'favors', 'of', 'your', 'lord', 'would', 'you', 'deny', '?', '\"']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'states', 'reported', 'tests', '40k', 'cases', 'and', '1202', 'deaths', '.', 'added', '7-day', 'averages', 'to', 'nationwide', 'overview', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'claiming', 'bodies', 'packed', 'in', 'body', 'bags', 'are', 'being', 'dumped', 'in', 'mass', 'graves', 'in', 'italy', 'and', 'spain', '.']\n",
            "pred:  0 label:  0 ['indian', 'army', 'has', 'built', 'a', 'hospital', 'in', 'rajasthan', 'in', 'eight', 'days', 'for', 'treatment', 'of', '.']\n",
            "pred:  0 label:  0 ['there', 'has', 'been', 'a', 'pandemic', 'every', '100', 'years']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'florida', 'hospital', 'reports', 'a', 'coronavirus', '?', '?', '?', 'infestation', '?', '?', '?', 'with', 'multiple', 'confirmed', 'patients', '.', '?', '?']\n",
            "pred:  0 label:  0 ['muslim', 'doctor', 'died', 'while', 'treating', 'people', 'infected', 'with', 'coronavirus', 'in', 'delhi']\n",
            "pred:  0 label:  0 ['as', 'tuberculosis', 'shaped', 'modernism', ',', 'so', 'and', 'our', 'collective', 'experience', 'of', 'staying', 'inside', 'for', 'months', 'on', 'end', 'will', 'influence', 'near', 'future', ',', 'writes', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'whatsapp', 'forward', 'has', 'gone', 'viral', 'which', 'claimed', 'that', 'a', 'vaccine', 'for', 'coronavirus', 'has', 'been', 'discovered', '.', 'viral', 'forward', 'also', 'claims', 'that', 'treatment', 'for', 'coronavirus', 'is', 'mentioned', 'in', 'intermediate', 'zoology', 'book', '.']\n",
            "pred:  1 label:  1 ['rt', 'a', 'vaccine', 'will', 'be', 'a', 'precious', 'resource', '.', 'unless', 'we', 'have', 'an', 'international', 'plan', 'to', 'manage', 'it', 'fairly', 'there', 'will', 'be']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'states', 'reported', 'tests', 'cases', 'and', '406', 'deaths', '.', 'number', 'of', 'cases', 'reported', 'is', 'lowest', 'since', 'june', '23', '.', '.']\n",
            "pred:  0 label:  0 ['several', 'social', 'media', 'users', 'claimed', 'that', 'there', 'was', 'a', 'massive', 'crowd', 'at', 'kr', 'market', 'on', 'occasion', 'of', 'flouting', 'social', 'distancing', 'norms', '.', 'this', 'claim', 'is', 'misleading', '.', 'a', '2019', 'video', 'is', 'being', 'passed', 'off', 'as', 'a', 'recent', 'one', '.', '.']\n",
            "pred:  1 label:  1 ['this', 'rise', 'in', 'deaths', 'is', 'concentrated', 'in', 'states', 'with', 'large', 'outbreaks', '.', 'texas', 'california', 'and', 'florida', 'all', 'reported', 'their', 'single', 'highest', 'day', 'of', 'deaths', 'for', 'entire', 'pandemic', 'today', '.']\n",
            "pred:  0 label:  0 ['will', 'do', 'nothing', 'for', 'coronavirus', '.', '\"']\n",
            "pred:  0 label:  0 ['coronavirus', 'is', 'only', 'as', 'dangerous', 'as', 'a', 'flu', 'diseases', 'have', 'same', 'death', 'rate', 'air', 'pollution', 'and', 'bad', 'hygienic', 'conditions', 'in', 'italian', 'hospitals', 'are', 'responsible', 'for', 'many', 'death', 'cases', 'and', 'only', 'of', 'whole', 'population', 'is', 'endangered', 'by', '.']\n",
            "pred:  0 label:  0 ['lockdown', 'has', 'now', 'lasted', 'so', 'long', 'children', 'could', 'have', 'put', 'their', 'shoes', 'on', 'twice', '.', '.']\n",
            "pred:  0 label:  0 ['quotes', 'joe', 'biden', 'as', 'saying', 'who', 'have', 'never', 'died', 'before', 'are', 'now', 'dying', 'from', 'coronavirus', '.', '”']\n",
            "pred:  1 label:  1 ['total', 'number', 'of', 'confirmed', 'cases', 'of', 'is', 'now', '1200', 'which', 'is', 'number', 'we', 'report', 'to', 'world', 'health', 'organization', '.', 'there', 'is', 'no', 'one', 'in', 'new', 'zealand', 'receiving', 'care', 'for', '.']\n",
            "pred:  0 label:  0 ['approval', 'from', 'ayush', '.', 'time', 'for', 'particular', 'lobby', 'to', 'buy']\n",
            "pred:  1 label:  1 ['recoveries', 'exceed', 'active', 'cases', 'by', 'more', 'than', '2', '.', '1', 'million', '.', 'there', 'has', 'been', 'more', 'than', '4', 'times', 'jump', 'in', 'average', 'weekly', 'recoveries', 'from', 'first', 'week', 'of', 'july', 'to', 'last', 'week', 'of', 'august', '.', '.']\n",
            "pred:  0 label:  0 ['jamaat', 'chief', 'maulana', '70000', 'people', 'get', 'coronavirus', 'india', 'will', 'be', 'destroyed', '.', 'country', 'will', 'be', 'in', 'our', 'possession', '.', '\"']\n",
            "pred:  1 label:  1 ['as', 'outbreak', 'surges', 'there', 'have', 'been', 'a', 'lot', 'of', 'questions', 'about', 'data', '.', 'a', 'very', 'deep', 'dive', 'by', 'and', 'into', 'what', 'we', 'know', 'wrong', 'and', 'missing', '.', '.', '.']\n",
            "pred:  0 label:  0 ['maine', 'governor', 'mandates', 'dog', 'cone', 'wearing', 'for', 'restaurant', 'staff', 'instead', 'of', 'mask', '.']\n",
            "pred:  1 label:  1 ['your', 'household', 'plan', 'for', 'should', 'include', 'talking', 'with', 'people', 'who', 'need', 'to', 'be', 'included', 'in', 'your', 'plan', ';', 'planning', 'ways', 'to', 'care', 'for', 'those', 'who', 'might', 'be', 'at', 'greater', 'risk', 'for', 'serious', 'complications', '.', 'learn', '.', '.']\n",
            "pred:  1 label:  1 ['india', 'registers', 'a', 'record', 'of', 'highest', 'single', 'day', 'recoveries', 'more', 'than', '1', 'lakh', 'patients', 'recovered', 'in', 'last', '24', 'hours', '.']\n",
            "pred:  1 label:  1 ['as', 'at', '08:00', 'pm', '31st', 'march', 'there', 'are', '139', 'confirmed', 'cases', '9', 'discharged', '2', 'deaths', 'for', 'a', 'breakdown', 'of', 'cases', 'by', 'states', 'in', 'real', 'time', 'visit', '.', 'currently', ';', '82', '28', '8', '5', '4', '3', '2', '2', '2', '1', '1']\n",
            "pred:  0 label:  0 ['photo', 'shows', 'last', 'meeting', 'of', 'a', 'turkish', 'doctor', 'who', 'died', 'due', 'to', 'with', 'his', 'child', 'in', 'munich', '.']\n",
            "pred:  1 label:  1 ['one', 'state', 'situation', 'worth', 'highlighting', 'is', 'michigan', '.', 'large', 'numbers', 'of', 'positives', 'deaths', 'rising', 'and', 'an', 'uncertain', 'testing', 'situation', '.', 'very', 'worrisome', 'overall', '.', '.']\n",
            "pred:  1 label:  1 ['only', 'aggressive', 'action', 'combined', 'with', 'national', 'unity', 'and', 'global', 'solidarity', 'can', 'turn', 'pandemic']\n",
            "pred:  1 label:  1 ['visit', '.', 'to', 'stay', 'up-to-date', 'on', 'latest', 'data', 'for', 'your', 'state', 'or', 'county', '.', 'our', 'warning', 'system', 'focuses', 'on', '4', 'key', 'infection', 'rate', 'positive', 'test', 'rate', 'icu', 'headroom', 'used', 'contact', 'traced', 'data', 'updates', 'every', '3', 'days', 'so', 'be', 'sure', 'to', 'check', 'back', 'regularly', '.']\n",
            "pred:  0 label:  0 ['18', 'states', 'including', 'us', 'uk', 'and', 'australia', 'request', 'pm', 'modi', 'to', 'head', 'a', 'task', 'force', 'to', 'stop', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['yesterday', 'our', 'laboratories', 'completed', 'tests', '.', 'that', 'brings', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  0 label:  0 ['we', 'a', 'range', 'of', 'statements', 'from', 'second', 'night', 'of', '.', 'here', 'are', 'highlights', 'dealing', 'with', 'coronavirus', 'case', 'numbers', 'and', 'position', 'on', '.', '.']\n",
            "pred:  0 label:  0 ['white', 'house', 'launches', 'coronavirus', 'medicine', '.']\n",
            "pred:  0 label:  0 ['video', 'of', 'pilgrims', 'who', 'were', 'moving', 'via', 'trucks', 'to', 'spread', 'in', 'india', '.']\n",
            "pred:  1 label:  1 ['a', 'approach', 'to', 'complex', 'coronary', 'bifurcation', 'lesions', 'yielded', 'improved', 'clinical', 'outcomes', '.', '.']\n",
            "pred:  0 label:  0 ['beware', 'of', 'a', 'donald', 'trump', 'vaccine', '.']\n",
            "pred:  1 label:  1 ['solidarity', 'is', 'needed', 'to', 'provide', 'a', 'joint', 'solution', 'to', 'pandemic', '.', 'vaccines', 'facility', 'is', 'critical', 'mechanism', 'for', 'joint', 'procurement', ';', 'pooling', 'risk', 'across', 'multiple', 'vaccines', 'which', 'is', 'why', 'i', 'sent', 'a', 'to', 'every', 'member', 'state', 'encouraging', 'them', 'to']\n",
            "pred:  1 label:  1 ['india', 'has', 'reported', 'highest', 'number', 'of', 'total', 'recoveries', 'with', 'more', 'than', '42', 'lakh', '.', '.']\n",
            "pred:  1 label:  1 ['testing', 'status', 'stated', 'that', 'samples', 'tested', 'upto', 'september', '15', '2020', 'sample', 'tested', 'on', 'september', '15', '2020', '.']\n",
            "pred:  1 label:  1 ['rt', 'india', 'crosses', '1', 'million', 'landmark', 'of', 'tests', 'in', '24', 'hours', '.', '.']\n",
            "pred:  0 label:  0 ['vaccine', 'for', 'new', 'coronavirus', 'has', 'been', 'created', 'by', 'us', 'scientists', 'and', 'will', 'be', 'ready', 'to', 'use', 'from', 'next', 'sunday', '.', '_']\n",
            "pred:  0 label:  0 ['not', 'for', 'but', 'to', 'prevent', 'other', 'infections', 'after', 'age', 'of', '65', 'consult', 'your', 'doctors', 'for', 'pneumococcal', 'conjugate', 'vaccine', 'or', 'pneumococcal', 'polysaccharide', 'vaccine', 'or', 'i', 'found', 'them', 'useful', 'for', 'my', 'wife', '.']\n",
            "pred:  1 label:  1 ['access', 'management', 'assessment', 'and', 'response', 'tool', '.', '.']\n",
            "pred:  0 label:  0 ['lockdown', '2020', 'survival', 'juice', 'get', 'yours', 'here', ';', ';', ';', '.', '.']\n",
            "pred:  1 label:  1 ['yesterday', 'our', 'laboratories', 'completed', '1007', 'tests', '.', 'seven', 'day', 'rolling', 'average', 'is', '1984', '.', 'that', 'brings', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  1 label:  1 ['rt', 'has', 'led', 'to', 'services', '–', 'have', 'you', 'utilized', 'these', 'virtual', 'appointments', '?', 'visit', '.']\n",
            "pred:  1 label:  1 ['states', 'reported', 'over', '50k', 'cases', 'for', 'first', 'time', 'today', 'another', 'new', 'record', '.', 'national', 'positivity', 'rate', 'is', 'over', 'now', 'despite', 'ever', 'higher', 'testing', 'levels', '.', 'hospitalizations', 'are', 'now', 'at', 'levels', '.', '.']\n",
            "pred:  1 label:  1 ['ages', '10-19', 'had', 'highest', 'rate', 'of', 'transmission', '.', '.', 'via', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'you', 'see', 'suicide', 'rates', 'are', 'skyrocketing', 'now', '?', '?', '?', '?', '?']\n",
            "pred:  1 label:  1 ['as', 'on', '25', 'july', '6pm', 'out', 'of', 'all', 'active', 'cases', '1', '.', 'are', 'in', '0', '.', 'are', 'on', 'and', '2', '.', 'require', '.']\n",
            "pred:  0 label:  0 ['why', 'censor', 'her', '?', 'dr', '.', 'li', 'meng', 'yan', 'claims', 'was', 'bio-engineered', 'by', 'red', 'c', '.', '.', '.', '.', 'via']\n",
            "pred:  0 label:  0 ['drinking', 'tea', 'and', 'hot', 'water', 'kills', 'virus', ';', 'lemon', 'slices', 'in', 'warm', 'water', 'may', 'save', 'your', 'life', ';', 'recommended', 'to', 'keep', 'your', 'money', 'in', 'a', 'sealed', 'plastic', 'bag', ';', 'a', 'conspiracy', 'theory', 'suggests', 'china', 'is', 'benefiting', 'from', 'pandemic', ';', 'vinegar', 'is', 'good', 'for', 'cleaning', 'your', 'hands', '.']\n",
            "pred:  0 label:  0 ['can', 'not', 'buy', 'anything', 'with', 'are', 'totally', 'chinese', 'yuan', 'becomes', 'just', 'merely', 'paper', 'and', 'trash', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'tweet', 'from', 'president', 'donald', 'trump', 'that', 'says', ',', 'before', 'u', '.', 's', '.', 'issues', 'financial', 'assistance', 'to', 'individuals', 'to', 'deal', 'with', 'coronavirus', 'pandemic', ',', 'will', 'go', 'through', 'your', 'social', 'media', 'history', 'from', 'last', '4', 'years', 'and', 'search', 'for', 'any', 'post', 'with', '.', '.', '.']\n",
            "pred:  1 label:  1 ['there', 'are', 'nine', 'previously', 'reported', 'cases', 'that', 'are', 'now', 'considered', 'to', 'have', 'recovered', 'from', '.', 'our', 'total', 'number', 'of', 'active', 'cases', 'is', '62', '-', 'of', 'those', '29', 'are', 'imported', 'cases', 'in', 'facilities', 'and', '33', 'are', 'community', 'cases', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'claims', 'that', 'bill', 'gates', 'made', 'a', 'presentation', 'to', 'cia', 'on', 'vaccines', 'for', 'modifying', 'brain', 'of', 'religious', 'fanatics', '.']\n",
            "pred:  1 label:  1 ['social', 'gatherings', 'of', 'more', 'than', 'six', 'people', 'will', 'be', 'banned', 'across', 'england', 'wales', 'and', 'scotland', 'from', 'tomorrow', '.', 'but', 'what', 'are', 'new', 'rules', 'what', 'happens', 'if', 'you', 'break', 'them', 'and', 'how', 'do', 'they', 'differ', 'across', 'nations', '?']\n",
            "pred:  1 label:  1 ['in', 'texas', 'too', 'daily', 'average', 'deaths', 'have', 'risen', 'in', 'last', 'two', 'weeks', '.', 'read', 'more', 'on', 'complexities', 'of', 'lag', 'times', 'changing', 'demographics', 'and', 'deaths', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'to', 'get', 'under', 'control', 'in', 'us', 'we', 'need', 'to', 'all', 'wear', 'face', 'masks', 'social', 'distance', 'and', 'wash', 'hands', 'frequently', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '24', 'million', 'tests', 'up', 'from', 'yesterday', '.', 'for', 'details', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'due', 'to', 'large', 'number', 'of', 'people', 'who', 'will', 'refuse', 'forthcoming', 'vaccine', 'because', 'it', 'will', 'include', 'tracking', 'microchips', ',', 'gates', 'foundation', 'is', 'now', 'spending', 'billions', 'to', 'ensure', 'that', 'all', 'medical', 'and', 'dental', 'injections', 'and', 'procedures', 'include', 'chips', '.', '?', '?']\n",
            "pred:  1 label:  1 ['we', 'reached', '3', 'million', 'cumulative', 'cases', 'today', 'less', 'than', 'a', 'month', 'after', 'reaching', '2', 'million', '.', '.']\n",
            "pred:  1 label:  1 ['you', 'can', 'help', 'slow', 'spread', 'of', '.', 'practice', '.', 'keep', 'at', 'least', 'six', 'feet', 'of', 'physical', 'distance', 'between', 'yourself', 'and', 'others', '.', 'wear', 'a', 'cloth', 'face', 'covering', 'when', 'in', 'public', '.', 'more', 'tips', 'at', '.', '.', '.']\n",
            "pred:  0 label:  0 ['audio', 'from', 'whatsapp', 'in', 'which', 'head', 'of', 'cardiology', 'at', 'gregorio', 'marañón', 'hospital', 'states', 'that', 'we', 'are', 'in', 'a', 'moment', 'of', 'uncontrolled', 'coronavirus', 'infection', '.']\n",
            "pred:  1 label:  1 ['prime', 'minister', 'boris', 'johnson', 'says', 'are', 'a', 'long', 'way', 'having', 'tests', 'and', 'adds', 'as', 'soon', 'as', 'those', 'tests', 'become', 'available', 'it', 'will', 'be', 'easier', 'for', 'theatres', 'and', 'football', 'stadiums', 'to', 'reopen', 'fully', '.', '.', '.']\n",
            "pred:  0 label:  0 ['man', 'has', 'gone', 'to', 'live', 'in', 'parallel', 'universe', '.']\n",
            "pred:  0 label:  0 ['alternative', 'medicine', 'guide', 'to', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['italy', '?', '?', '?', 'went', 'against', '?', '?', '?', 'law', '?', '?', '?', '?', '?', 'against', 'autopsies', 'and', '?', '?', '?', 'discovered', 'is', 'not', 'a', 'virus', ',', 'but', 'a', 'bacterium', '?', '?', 'that', '?', '?', '?', 'clots', 'blood', 'and', 'reduces', 'oxygen', 'saturation', '.', '?', '?']\n",
            "pred:  0 label:  0 ['post', 'claims', 'that', 'sister', 'of', 'bharatiya', 'janata', 'party', 'leader', 'kapil', 'mishra', 'married', 'a', 'muslim', 'man']\n",
            "pred:  1 label:  1 ['6', 'new', 'symptoms', 'have', 'been', 'added', 'to', 'list', 'of', 'symptoms', 'associated', 'with', '.', 'along', 'with', 'fever', 'cough', 'and', 'shortness', 'of', 'breathe', 'has', 'chills', 'repeated', 'shaking', 'with', 'chills', 'muscle', 'pain', 'headache', 'sore', 'throat', 'and', 'new', 'loss', 'of', 'taste', 'or', 'smell', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'post', 'shared', 'more', 'than', 'times', 'on', 'facebook', 'during', 'novel', 'coronavirus', 'pandemic', 'says', 'bill', 'gates', 'wants', 'to', 'check', 'who', 'has', 'been', 'tested', 'and', 'asks', 'if', 'it', 'would', 'be', 'holocaust', 'victims', 'have', '.', '”']\n",
            "pred:  1 label:  1 ['following', 'national', 'trend', '17', 'have', 'more', 'new', 'recoveries', 'than', 'new', 'cases', '.', '.']\n",
            "pred:  0 label:  0 ['meditate', 'your', 'way', 'through', '19', 'with', 'tracy', 'honeycomb', '.']\n",
            "pred:  0 label:  0 ['a', 'photo', 'of', 'people', 'lying', 'in', 'street', 'is', 'accompanied', 'by', 'a', 'caption', 'implying', 'they', 'are', 'coronavirus', 'victims', 'in', 'china', '.']\n",
            "pred:  1 label:  1 ['officials', 'are', 'skeptical', 'that', 'supply', 'address', 'demand', 'during', 'a', 'second', 'wave', 'of', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'viral', 'video', 'shows', 'a', 'group', 'of', 'doctors', 'airing', 'unproven', 'conspiracy', 'theories', 'about', 'coronavirus', '.', 'one', 'of', 'most', 'inaccurate', 'claims', 'comes', 'from', 'dr', '.', 'stella', 'immanuel', 'who', 'has', 'a', 'track', 'record', 'of', 'making', 'bizarre', 'medical', 'claims', 'such', 'as', 'believing', 'in', 'alien', 'dna', '.', '.']\n",
            "pred:  1 label:  1 ['schoolchildren', 'and', 'their', 'parents', 'could', 'be', 'prioritised', 'for', 'coronavirus', 'tests', '-', 'after', 'hospitals', 'and', 'care', 'homes', '-', 'as', 'government', 'deals', 'with', 'in', 'system', 'a', 'cabinet', 'minister', 'has', 'told', 'sky', 'news', '.']\n",
            "pred:  1 label:  1 ['review', '7', 'vaccines', 'currently', 'in', 'phase', '3', 'trials', '.', '.', 'via', '.']\n",
            "pred:  1 label:  1 ['of', 'total', 'active', 'cases', 'are', 'found', 'in', 'nine', 'most', 'affected', '.', 'new', 'cases', 'have', 'been', 'reported', 'in', 'last', '24', 'hours', 'in', 'country', '.', 'details', ':', '.', '.']\n",
            "pred:  1 label:  1 ['several', 'sites', 'tracking', 'hit', 'a', 'grim', 'milestone', 'more', 'than', 'deaths', 'since', 'pandemic', 'began', '.', 'our', 'figures', 'yet', 'reached', 'that', 'level', '.', 'why', '.', '.']\n",
            "pred:  1 label:  1 ['there', 'are', '4', 'govt', '.', 'testing', 'lab', 'in', '.', 'kindly', 'refer', 'to', 'following', 'link', 'for', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'joe', 'biden', 'and', 'democrats', 'want', 'to', 'prosecute', 'americans', 'for', 'going', 'to', 'church', ',', 'but', 'not', 'for', 'burning', 'a', 'church', '.', '?', '?']\n",
            "pred:  0 label:  0 ['dr', '.', 'anthony', 'fauci', 'known', 'for', '15', 'years', 'that', 'chloroquine', 'and', '?', '?', '?', 'will', 'not', 'only', 'treat', 'a', 'current', 'case', 'of', 'coronavirus', 'but', 'prevent', 'future', 'cases', '.', '?', '?']\n",
            "pred:  0 label:  0 ['there', 'are', '?', '?', '?', '360', ',', '000', 'a', 'year', 'from', 'swimming', 'pools', '.', 'but', 'we', 'don', '?', '?', '?', 't', 'shut', 'country', 'down', 'for', 'that', '.', '?', '?']\n",
            "pred:  0 label:  0 ['an', 'inmate', 'who', 'was', 'released', 'from', 'prison', 'under', 'house', 'was', 'arrested', 'again', 'in', 'campo', 'bom', '.', 'police', 'officers', 'found', 'more', 'than', '100', 'kg', 'of', 'cocaine', 'in', 'his', 'house', 'and', 'several', 'guns', '.']\n",
            "pred:  1 label:  1 ['also', 'like', 'to', 'call', 'attention', 'to', 'incredible', 'job', 'florida', 'and', 'are', 'doing', 'releasing', 'data', '.', 'not', 'only', 'do', 'they', 'provide', 'hospitalization', 'rates', 'and', 'testing', 'results', 'but', 'a', 'ton', 'of', 'other', '.']\n",
            "pred:  1 label:  1 ['one', 'of', 'lowest', 'globally', 'case', 'fatality', 'rate', '1', '.', 'and', 'sliding', 'health', 'ministry', 'issues', 'faqs', 'on', 'better', 'clinical', 'management', 'of', 'patients', 'in', 'icus', '.']\n",
            "pred:  0 label:  0 ['a', 'person', 'with', 'a', 'cold', 'or', 'flu', 'will', 'test', 'positive', 'for']\n",
            "pred:  0 label:  0 ['of', 'part', 'of', 'stimulus', 'will', 'mobilise', 'equity', 'of', '50000', 'cr', 'for', 'a', 'similar', 'fund', 'set', 'up', 'in', '2016', 'to', 'mobilise', 'rs', '60000', 'cr', 'for', 'start-ups', 'disbursed', 'only', 'of', 'planned', 'amount', '.']\n",
            "pred:  0 label:  0 ['detainees', 'in', 'central', 'prison', 'burundi', 'died', 'due', 'to', 'and', 'that', 'some', 'contracted', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['now', 'moving', 'to', 'a', 'new', 'phase', 'where', 'many', 'of', 'us', 'are', 'heading', 'back', 'to', 'work', 'and', 'school', 'for', 'first', 'time', 'in', 'a', 'while', '.', 'stay', 'healthy', 'and', 'stay', 'safe', 'new', 'zealand', '.']\n",
            "pred:  0 label:  0 ['bechara', 'se', 'down', 'hai', 'else', 'he', 'would', 'have', 'give', 'a', 'nice', 'acceptance', 'speech', '.', 'dad', 'this', 'is', 'for', 'you', '.', '.']\n",
            "pred:  0 label:  0 ['says', 'coronavirus', 'was', 'engineered', 'by', 'scientists', 'in', 'a', 'lab', '.', '\"']\n",
            "pred:  1 label:  1 ['arizona', 'reported', 'more', 'cases', 'today', 'than', 'on', 'any', 'previous', 'single', 'day', '.', '.']\n",
            "pred:  1 label:  1 ['prime', 'minister', 'has', 'put', 'his', 'faith', 'in', 'operation', 'moonshot', 'but', 'meanwhile', 'on', 'planet', 'earth', 'there', 'was', 'no', 'nhs', 'tests', 'available', 'for', 'several', 'high', 'infection', 'areas', 'says', 'labour', 'deputy', 'leader', 'angela', 'rayner', 'at', '.', 'follow', 'live', '.', '.']\n",
            "pred:  0 label:  0 ['video', 'of', 'a', 'doctor', 'at', 'a', 'press', 'conference', 'who', 'said']\n",
            "pred:  0 label:  0 ['5g', 'is', 'cause', 'of', 'coronavirus', 'pandemic', '.']\n",
            "pred:  0 label:  0 ['it', 'effect', 'coz', 'they', 'have', 'for', 'after', 'all', 'they', 'have', 'bad', 'smell', 'to', 'tackle', 'with']\n",
            "pred:  1 label:  1 ['only', 'way', 'to', 'tell', 'if', 'a', 'kid', 'just', 'has', 'a', 'cold', 'can', 'go', 'to', 'is', 'robust', 'rapid', 'testing', 'for', '—', 'and', 'right', 'now', 'we', 'have', 'it', '.', '.', 'via', '.']\n",
            "pred:  1 label:  1 ['at', 'time', 'they', 'did', 'not', 'meet', 'case', 'definition', 'for', 'testing', 'for', 'people', 'with', 'defined', 'symptoms', 'who', 'had', 'travelled', 'from', 'or', 'transited', 'through', 'china', '.', 'italy', 'had', 'not', 'at', 'that', 'point', 'been', 'identified', 'as', 'a', 'country', 'of', 'concern', '.', 'this', 'meant', 'new', 'zealand', 'household', 'was', 'not', 'tested', 'at', 'time', '.']\n",
            "pred:  1 label:  1 ['many', 'states', 'reported', 'a', 'huge', 'number', 'of', 'tests', 'including', 'a', 'known', 'backlog', 'clearing', 'from', 'ma', '.', 'ny', 'reported', 'almost', '50k', 'tests', ';', 'tx', 'over', '20k', '.', 'al', 'ca', 'fl', 'ga', 'il', 'ma', 'ny', 'tn', 'tx', 'all', 'reported', 'over', '10k', 'tests', '.']\n",
            "pred:  1 label:  1 ['some', 'unexpected', 'numbers', 'came', 'from', 'indiana', 'today', '.', 'despite', 'only', 'having', 'confirmed', 'cases', 'state', 'said', 'today', 'it', 'has', '924', 'patients', 'in', 'icu', 'and', '500', 'on', 'ventilators', '.', 'not', 'too', 'far', 'from', 'icu', 'count', 'and', 'closing', 'in', 'on', 'pennsylvania', 'and', 'vent', 'usage', '.']\n",
            "pred:  0 label:  0 ['opening', 'ceremony', 'of', '2012', 'london', 'olympics', 'was', 'a', 'projection', 'of', 'coronavirus', 'pandemic', '.', 'hospital', 'beds', 'on', 'scene', 'during', 'opening', 'show', 'and', 'boris', 'johnson', 'in', 'a', 'hospital', 'bed', 'prove', 'that', 'spread', 'of', 'disease', 'was', 'planned', 'in', 'advance', '.']\n",
            "pred:  1 label:  1 ['transmission', 'may', 'include', 'viral', 'particles', 'that', 'remain', 'airborne', 'for', 'longer', 'than', 'previously', 'understood', '.', '.']\n",
            "pred:  0 label:  0 ['who', 'banned', 'autopsies', 'and', 'italy', 'changed', 'protocols', 'for', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'according', 'to', 'cdc', ',', 'so', 'far', 'this', 'year', ',', 'florida', 'has', 'had', '1', ',', '762', 'deaths', 'from', 'and', '5', ',', '185', 'from', 'pneumonia', '.', 'average', 'pneumonia', 'deaths', 'in', 'florida', 'from', 'for', 'same', 'time', 'period', 'are', '918', '.', '?', '?']\n",
            "pred:  1 label:  1 ['there', 'were', 'an', 'estimated', '6000', 'new', 'cases', 'every', 'day', 'in', 'england', 'during', 'first', 'week', 'of', 'september', '.']\n",
            "pred:  1 label:  1 ['reported', 'deaths', 'rose', 'to', '1430', 'today', '.', 'important', 'to', 'recognize', 'that', 'this', 'data', '.', '.']\n",
            "pred:  0 label:  0 ['underlying', 'cause', 'of', 'death', 'in', 'vast', 'majority', 'of', 'death', 'certificates', 'that', 'mention', 'is', 'coronavirus', '.', '.']\n",
            "pred:  1 label:  1 ['a', 'critical', 'lack', 'of', 'representative', 'clinical', 'images', 'of', 'skin', 'manifestations', 'in', 'patients', 'of', 'color', '.', '.']\n",
            "pred:  1 label:  1 ['injecting', 'genetic', 'material', 'into', 'host', 'so', 'that', 'host', 'cells', 'create', 'proteins', 'that', 'are', 'similar', 'to', 'those', 'in', 'virus', 'against', 'which', 'host', 'then', 'creates', 'antibodies']\n",
            "pred:  0 label:  0 ['claim', 'stated', 'that', 'top', 'business', 'conglomerate', 'tata', 'group', 'chairman', 'ratan', 'tata', 'said', 'not', 'time', 'to', 'think', 'of', 'profits', 'but', 'to', 'think', 'of', 'survival', '.']\n",
            "pred:  1 label:  1 ['our', 'weekly', 'update', 'is', 'posted', 'and', 'finally', 'some', 'good', 'nationally', 'cases', 'are', 'declining', 'and', 'states', 'with', 'worst', 'outbreaks', 'seem', 'to', 'be', 'turning', 'a', '.']\n",
            "pred:  0 label:  0 ['remedy', 'for', 'corona', 'virus', 'may', 'have', 'existed', 'for', 'a', 'long', 'time', ',', '1934', 'old', 'malaria', 'drug', 'seems', 'to', 'work', 'against', 'coronavirus', 'symptoms', '.', 'made', 'in', 'holland']\n",
            "pred:  0 label:  0 ['madrid', 'has', 'enabled', 'phone', 'number', 'to', 'request', 'prescriptions', '.']\n",
            "pred:  0 label:  0 ['claim', 'that', 'govt', 'is', 'providing', 'money', 'to', 'bereaved', 'families', 'of', 'victims', 'is', 'false']\n",
            "pred:  1 label:  1 ['there', 'are', 'nine', 'people', 'in', 'hospital', '-', '2', 'in', 'auckland', 'city', 'three', 'in', 'north', 'shore', 'in', 'and', 'rest', 'in', 'middlemore', '2', 'are', 'in', 'icu', 'and', 'considered', 'critical', '.', 'we', 'understand', 'that', 'this', 'is', 'a', 'very', 'difficult', 'time', 'for', 'them', 'and', 'their', 'families', '.', 'we', 'are', 'wishing', 'them', 'a', 'full', 'and', 'speedy', 'recovery', '.']\n",
            "pred:  1 label:  1 ['india', 'tracker', 'on', '23', 'september', '2020', '08:00', '.', '.', '.', 'via', '.']\n",
            "pred:  1 label:  1 ['216', 'new', 'cases', 'of', ';', 'cases', 'of', 'in', 'nigeria', '1644', '191', '.']\n",
            "pred:  1 label:  1 ['a', 'christchurch', 'person', 'discussed', 'yesterday', 'as', 'being', 'under', 'further', 'investigation', 'has', 'now', 'been', 'confirmed', 'as', 'not', 'a', 'case', 'of', '.', 'investigation', 'has', 'been', 'closed', '.']\n",
            "pred:  0 label:  0 ['previously', 'they', 'claimed', 'that', 'cow', 'urine', 'cures', 'qnd', 'people', 'actually', 'believed', 'it', '.', '.']\n",
            "pred:  1 label:  1 ['number', 'of', 'daily', 'tests', 'has', 'been', 'increasing', 'in', 'a', 'steep', 'climb', '.', 'average', 'daily', 'tests', 'during', 'past', 'three', 'weeks', 'also', 'strongly', 'depict', 'progress', 'made', 'in', 'enhancement', 'of', 'tests', 'across', 'country', '.', '.']\n",
            "pred:  0 label:  0 ['tracing', 'apps', 'have', 'been', 'secretly', 'installed', 'on', 'every', 'android']\n",
            "pred:  0 label:  0 ['table', 'that', 'compares', 'deaths', 'from', 'with', 'other', 'causes', 'show', 'new', 'coronavirus', 'is', 'harmless', '.']\n",
            "pred:  1 label:  1 ['watch', 'cdc', 'experts', 'discuss', 'managing', 'chronic', 'illness', 'during', 'pandemic', 'in', 'this', 'facebook', 'live', 'hosted', 'by', 'and', 'global', 'healthy', 'living', 'foundation', '.', '.']\n",
            "pred:  0 label:  0 ['man', 'finds', 'it', 'difficult', 'to', 'eat', 'when', 'wearing', 'a', 'face', 'mask', '.']\n",
            "pred:  1 label:  1 ['virtual', 'rounding', 'clinical', 'exams', 'and', 'lectures', '.', 'may', 'permanently', 'change', '.', '.']\n",
            "pred:  0 label:  0 ['claim', 'that', 'indian', 'prime', 'minister', 'modi', 'said', 'one', 'crore', 'infected', 'patients', 'have', 'been', 'treated', 'for', 'free', '.']\n",
            "pred:  1 label:  1 ['case', 'in', 'managed', 'isolation', 'is', 'a', 'child', 'who', 'arrived', 'in', 'nz', 'from', 'afghanistan', 'via', 'dubai', '.', 'person', 'has', 'been', 'in', 'managed', 'isolation', 'at', 'pullman', '.', 'person', 'tested', 'positive', 'following', 'day', '12', 'routine', 'testing', '.']\n",
            "pred:  1 label:  1 ['there', 'is', 'one', 'person', 'receiving', 'care', 'for', ';', 'they', 'are', 'in', 'middlemore', 'and', 'are', 'not', 'in', 'icu', '.', 'yesterday', 'our', 'laboratories', 'completed', 'tests', 'bringing', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'shows', 'that', 'gates', 'admits', 'vaccine', 'will', 'no', 'doubt', 'kill', '700', ',', '000']\n",
            "pred:  0 label:  0 ['cattle', 'vaccine', 'proofs', 'that', 'coronavirus', 'has', 'existed', 'for', 'years', 'and', 'that', 'there', 'already', 'is', 'a', 'vaccine', '.', 'photo', 'showing', 'cattle', 'vaccine', 'produced', 'by', 'american', 'medical', 'company', '.']\n",
            "pred:  0 label:  0 ['man', 'did', 'pass', 'along', 'coronavirus', 'infection', 'at', 'a', 'walmart', 'in', 'louisiana', '.']\n",
            "pred:  1 label:  1 ['an', 'even', 'better', 'piece', 'of', 'states', 'reported', 'fewer', 'than', '500', 'deaths', '.', 'that', 'happened', 'since', 'march', '.', 'yes', 'there', 'will', 'probably', 'be', 'a', 'larger', 'number', 'of', 'deaths', 'reported', 'tomorrow', 'as', 'lagging', 'weekend', 'data', 'gets', 'posted', 'but', 'it', 'is', 'a', 'significant', 'pandemic', 'milestone', '.', '.']\n",
            "pred:  0 label:  0 ['novel', 'coronavirus', 'not', 'last', 'in', 'philippines', 'because', 'of', 'its', 'warm', 'climate', '.']\n",
            "pred:  0 label:  0 ['lockdown', 'forcing', 'woman', 'to', 'realise', 'just', 'how', 'bad', 'her', 'diet', 'is', '.', '.']\n",
            "pred:  0 label:  0 ['in', 'future', '-', 'our', 'world', ',', 'some', 'hollywood', 'blockbuster', 'movies', 'are', 'expected', 'to', 'earn', 'as', 'much', 'as', 'in', 'ticket']\n",
            "pred:  1 label:  1 ['if', 'we', 'look', 'at', 'critical', 'cases', 'across', 'country', 'we', 'find', 'that', '.', 'of', 'total', 'actives', 'are', 'on', 'oxygen', 'support', '.', 'are', 'in', 'icu', '0', '.', 'patients', 'are', 'on', 'ventilator', 'secretary', '.']\n",
            "pred:  1 label:  1 ['india', 'reaches', 'another', 'record', 'of', 'highest', 'single', 'day', 'recoveries', 'at', '.', '.']\n",
            "pred:  1 label:  1 ['two', 'more', 'grim', 'milestones', '*', 'over', '1000', 'people', 'in', 'u', '.', 's', '.', 'have', 'died', '*', '10000', 'people', 'have', 'been', 'hospitalized', 'at', 'some', 'point', '.', 'seeing', 'more', 'than', '80000', 'total', 'positive', 'tests', 'in', 'our', 'data', '.']\n",
            "pred:  0 label:  0 ['30-year-old', 'indian', 'textbook', 'lists', 'aspirin', ',', 'and', 'nasal', 'spray', 'as', 'treatments', 'for', '.']\n",
            "pred:  1 label:  1 ['this', 'ensuring', 'that', 'no', 'one', 'leaves', 'a', 'managed', 'isolation', 'facility', 'without', 'having', 'had', 'a', 'negative', 'test', 'ensuring', 'that', 'all', 'people', 'in', 'isolation', 'are', 'tested', 'on', 'around', 'days', '3', 'and', '12', '.']\n",
            "pred:  0 label:  0 ['if', 'it', 'is', 'true', '.', '.', 'responsible', 'for', 'recent', '19', 'position', '.', '.', 'sack', 'those', 'corrupt', '.', '.', '.']\n",
            "pred:  1 label:  1 ['252', 'new', 'cases', 'of', ';', 'confirmed', 'discharged', '1007', 'deaths', '.']\n",
            "pred:  0 label:  0 ['man', 'reads', 'purple', 'by', 'm', '.', 'p', '.', 'shiel', ',', 'and', 'starts', 'to', 'get', 'a', 'bit', 'frightened', '.']\n",
            "pred:  1 label:  1 ['we', 'have', '1447', 'people', 'reported', 'as', 'having', 'recovered', 'from', '5', 'more', 'people', 'than', 'yesterday', '.', 'this', 'is', 'of', 'all', 'cases', '.', 'there', 'is', 'just', 'one', 'person', 'receiving', 'care', 'for', '.', 'they', 'are', 'not', 'in', 'icu', '.', 'we', 'still', 'have', '16', 'significant', 'clusters', '.']\n",
            "pred:  1 label:  1 ['yesterday', 'laboratories', 'processed', 'tests', 'for', 'bringing', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '.', 'laboratories', 'are', 'now', 'processing', 'tests', 'within', '24', 'hours', 'of', 'receiving', 'them', 'although', 'people', 'should', 'allow', 'four', 'days', 'to', 'be', 'notified', 'of', 'their', 'test', 'results', '.']\n",
            "pred:  1 label:  1 ['however', 'corticosteroids', 'do', 'not', 'help', 'patients', 'with', 'mild', 'or', 'moderate', 'disease', 'and', 'can', 'be', 'harmful', '.', 'who', 'therefore', 'recommends', 'use', 'of', 'corticosteroids', 'only', 'in', 'patients', 'who', 'are', 'severely', 'or', 'critically', '.']\n",
            "pred:  1 label:  1 ['fauci', 'says', 'rushing', 'out', 'vaccine', 'could', 'jeopardize', 'testing', 'of', 'others', '.', '.', '.']\n",
            "pred:  1 label:  1 ['342', 'people', 'have', 'been', 'referred', 'for', 'a', 'test', 'that', 'we', 'are', 'awaiting', 'results', 'for', '.']\n",
            "pred:  0 label:  0 ['is', 'a', 'man', 'made', 'virus', 'according', 'to', 'video', '.']\n",
            "pred:  0 label:  0 ['video', 'purports', 'to', 'show', 'a', 'police', 'officer', 'being', 'beaten', 'has', 'been', 'viewed', 'tens', 'of', 'thousands', 'of', 'times', 'on', 'facebook', 'and', 'twitter', 'alongside', 'a', 'claim', 'that', 'assault', 'at', 'a', 'temple', 'in', 'india', 'was', 'sparked', 'by', 'attempt', 'to', 'enforce', 'a', 'nationwide', 'novel', 'coronavirus', 'lockdown', '.', '_']\n",
            "pred:  0 label:  0 ['bill', 'gates', 'say', 'people', 'have', 'a', 'about', 'being', 'vaccinated', 'for', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['kylie', 'minogue', 'says', 'it', 'breaks', 'her', 'heart', 'that', 'coronavirus', 'has', 'denied', 'thousands', 'of', 'people', 'joy', 'of', 'seeing', 'this', 'glastonbury', 'festival', 'descend', 'into', 'a', 'hellish', 'mudbath', '.']\n",
            "pred:  1 label:  1 ['back', 'on', 'campus', '?', 'if', 'you', 'go', 'out', 'wear', 'a', 'mask', 'stay', '6', 'feet', 'apart', 'from', 'others', 'and', 'meet', 'in', 'outdoor', 'spaces', '.', 'learn', 'more', 'about', 'steps', 'you', 'can', 'take', 'to', 'help', 'protect', 'yourself', 'and', 'your', 'friends', 'from', '.', '.', '.']\n",
            "pred:  0 label:  0 ['first', 'city', 'in', 'world', 'conquers', 'with', 'chlorine', 'dioxide', '.', 'this', 'great', 'news', 'has', 'already', 'been', 'suppressed', 'by', 'dark', '.']\n",
            "pred:  0 label:  0 ['in', 'middle', 'of', 'a', 'global', 'pandemic', ',', 'trump', 'administration', 'is', 'still', 'working', 'to', 'gut', 'affordable', 'care', 'act', 'and', 'rip', 'health', 'care', 'away', 'from', 'millions', '.', 'morally', 'reprehensible', '.', 'they', 'need', 'to', 'drop', 'lawsuit', 'immediately', '.', '.']\n",
            "pred:  1 label:  1 ['latest', 'updates', 'from', 'remains', 'a', 'serious', 'pandemic', 'continuing', 'to', 'affect', 'many', 'countries', '.', 'we', 'are', 'not', 'immune', 'to', 'further', 'cases', 'arriving', 'on', 'our', 'shores', '.', '.']\n",
            "pred:  1 label:  1 ['as', 'of', 'today', 'of', 'cases', 'in', 'is', '68', '.', '.']\n",
            "pred:  0 label:  0 ['woman', 'burned', 'her', 'hands', 'while', 'putting', 'on', 'stove', 'due', 'to', 'high', 'alcohol', 'content', 'in', 'hand']\n",
            "pred:  1 label:  1 ['there', 'is', 'a', 'daily', 'number', 'you', 'can', 'get', 'at', 'any', 'time', '.', 'then', 'there', 'is', 'current', 'number', '.', 'we', 'tend', 'to', 'do', '3', 'runs', 'per', 'day', '.', 'big', 'midday', 'run', 'we', 'go', ';', 'then', 'runs', 'at', 'night', 'and', 'in', 'morning', '.']\n",
            "pred:  1 label:  1 ['we', 'always', 'appreciate', 'questions', 'about', 'quality', 'of', 'our', 'data', '.', 'if', 'you', 'see', 'a', 'number', 'that', 'look', 'right', 'please', 'file', 'an', 'issue', 'at', '.', 'and', 'we', 'will', 'investigate', '.']\n",
            "pred:  0 label:  0 ['there', 'is', 'a', 'nipah', 'virus', 'outbreak', 'in', 'india', 'amidst', 'coronavirus', 'pandemic', ',', 'according', 'to', 'a', 'warning', 'form', 'who', '.']\n",
            "pred:  1 label:  1 ['when', 'we', 'know', 'actual', 'number', 'is', 'much', 'higher', 'as', 'we', 'see', 'here', 'on', 'a', 'sonoma', 'county', 'dashboard', '.', '.', '.']\n",
            "pred:  0 label:  0 ['hates', 'women', 'leaders', ',', 'loves', 'trump', '.']\n",
            "pred:  0 label:  0 ['trump', 'says', 'he', 'will', 'nuke', 'china', 'if', 'he', 'knows', ',', 'for', 'sure', ',', 'they', 'are', 'serving', 'bat', 'soup', '.']\n",
            "pred:  1 label:  1 ['labour', 'leader', 'sir', 'keir', 'starmer', 'is', 'after', 'a', 'member', 'of', 'his', 'household', 'displayed', 'possible', 'symptoms', '.']\n",
            "pred:  0 label:  0 ['french', 'astrologer', 'nostradamus', 'had', 'predicted', 'in', '1551', 'that', 'there', 'would', 'be', 'an', 'outbreak', 'of', 'coronavirus', 'in', 'china', 'in', '2020', '.']\n",
            "pred:  1 label:  1 ['rt', 'five', 'states', '-', 'maharashtra', 'andhra', 'pradesh', 'karnataka', 'uttar', 'pradesh', 'and', 'tamil', 'nadu', 'account', 'for', 'of', 'active', 'cases', 'in']\n",
            "pred:  1 label:  1 ['rt', 'going', 'to', 'a', 'cooling', 'center', 'to', 'beat', 'this', '?', 'follow', 'these', 'tips', 'to', 'protect', 'yourself', 'from']\n",
            "pred:  0 label:  0 ['coronavirus', 'victims', 'are', 'buried', 'in', 'morning', 'this', 'wednesday', 'in', 'italy', '.', 'rest', 'in', 'peace', '.']\n",
            "pred:  0 label:  0 ['some', 'people', 'have', 'stopped', 'electricity', 'during', 'night', 'time', 'in', 'and', 'scattered', 'currency', 'notes', 'on', 'road', 'to', 'spread', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['pandemic', 'is', 'fake', '.']\n",
            "pred:  0 label:  0 ['hospital', 'of', 'gainesville', 'florida', 'has', 'confirmed', 'its', 'first', 'case', 'of', 'coronavirus', '.', '”']\n",
            "pred:  1 label:  1 ['death', 'toll', 'was', 'relatively', 'what', 'it', 'was', 'last', 'sunday', '.', 'in', 'past', 'weeks', 'there', 'has', 'tended', 'to', 'be', 'a', 'weekend', 'lag', 'effect', 'where', 'sunday', 'and', 'monday', 'numbers', 'are', 'lower', 'than', 'mid-week', 'days', '.', '.']\n",
            "pred:  0 label:  0 ['prince', 'andrew', 'to', 'officiate', 'at', 'grand', 'reopening', 'of', 'woking', 'pizza', 'express', '.']\n",
            "pred:  1 label:  1 ['testing', 'status', 'stated', 'that', 'samples', 'tested', 'upto', 'september', '22', '2020', 'sample', 'tested', 'on', 'september', '22', '2020', '.']\n",
            "pred:  0 label:  0 ['strange', 'coincidence', 'that', 'all', 'worst', 'affected', 'cases', 'are', 'along', 'same', 'latitude', ':', 'from', 'france', 'italy', 'iran', 'wuhan', 's', '.', 'korea', 'japan', 'seattle', 'washington', 'newyork', '.']\n",
            "pred:  1 label:  1 ['papua', 'new', 'guinea', 'now', 'has', 'supplies', 'to', 'test', 'for', 'in', 'all', 'provinces', 'thanks', 'to', 'a', 'strong', 'collaboration', 'between', 'who', 'australia', 'and', 'new', 'zealand', '.', 'more', 'details', '.', '.']\n",
            "pred:  0 label:  0 ['several', 'social', 'media', 'users', 'shared', 'a', 'picture', 'of', 'a', 'cow', 'and', 'a', 'leopard', 'and', 'claimed', 'that', 'it', 'was', 'taken', 'in', 'assam', '.', 'this', 'claim', 'is', 'misleading', '.', 'viral', 'image', 'has', 'nothing', 'to', 'do', 'with', 'assam', 'two', 'animals', 'spent', 'time', 'together', 'in', 'vadodara', 'in', '2002', '.', '.']\n",
            "pred:  0 label:  0 ['self', 'testing', 'by', 'breath', 'and', 'put', 'mustard', 'oil', 'through', 'nose', 'will', 'kill', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['our', 'discharges', 'today', 'include', '1325', 'community', 'recoveries', 'over', 'last', 'two', 'weeks', 'reported', 'from', 'lagos', 'state', 'in', 'line', 'with', 'new', 'case', 'management', 'guidelines', '.', 'a', 'breakdown', 'of', 'cases', 'by', 'state', 'can', 'be', 'found', 'via', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'when', 'using', 'a', 'medical', 'mask', 'you', '?', '?', '?', 're', 'supposed', 'to', 'use', '?', '?', '?', 'white', 'side', 'out', 'is', 'filter', 'for', 'when', 'you', '?', '?', '?', 're', 'not', 'sick', '.', '?', '?']\n",
            "pred:  0 label:  0 ['local', 'badass', 'only', 'washes', 'hands', 'for', '15', 'seconds', '.', '.']\n",
            "pred:  0 label:  0 ['cbs', 'aired', 'wrong', 'video', 'footage', 'of', 'a', 'coronavirus', 'hospital', '.']\n",
            "pred:  0 label:  0 ['just', 'casually', 'carrying', 'a', 'body', 'with', '1', 'hand', '.', '.', '?', 'sure', '.', '.', 'more', 'evidence', 'of', 'coronavirus', 'being', 'a', 'complete', 'hoax', 'with', 'props', '&', 'street', 'theater', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'states', 'reported', 'tests', 'cases', 'and', '475', 'deaths', '.', 'hospitalizations', 'are', 'at', 'their', 'lowest', 'level', 'since', 'june', '30', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'photo', 'that', 'has', 'gone', 'viral', 'in', 'social', 'media', 'shows', 'a', 'military', 'convoy', 'transporting', 'people', 'who', 'died', 'from', 'pandemic', 'in', 'italy', '.']\n",
            "pred:  1 label:  1 ['rt', 'based', 'on', 'what', 'we', 'currently', 'know', 'transmission', 'primarily', 'occurs', 'when', 'people', 'are', 'showing', 'symptoms', 'but', 'can', 'also', 'happen']\n",
            "pred:  0 label:  0 ['bill', 'gates', 'predicted', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['reported', 'deaths', 'were', 'also', 'low', 'today', 'just', '629', '.', 'if', 'these', 'numbers', 'keep', 'up', '7-day', 'average', 'will', 'fall', 'below', '1000', '.', '.']\n",
            "pred:  1 label:  1 ['although', 'we', 'still', 'face', 'many', 'challenges', 'who', 'is', 'encouraged', 'by', 'progress', 'we', 'have', 'made', 'against', 'this', 'new', 'virus', 'in', '8', 'months', '.', 'with', 'national', 'unity', 'and', 'global', 'solidarity', 'we', 'can', 'and', 'will', 'end']\n",
            "pred:  0 label:  0 ['a', 'police', 'officer', 'in', 'hajipur', 'jail', 'has', 'been', 'affected', 'with', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['who', 'is', 'in', 'final', 'stages', 'of', 'adding', 'more', 'partners', 'and', 'encourages', 'pharmaceutical', 'and', 'tech', 'companies', 'to', 'join', 'this', 'initiative', 'which', 'will', 'help', 'people', 'reduce', 'their', 'risk', 'of', 'and', 'lead', 'healthier', '.']\n",
            "pred:  1 label:  1 ['our', 'models', 'estimate', 'that', 'has', 'a', 'high', 'positive', 'test', 'rate', 'and', 'is', 'near', 'full', 'icu', 'capacity', '.', 'louisiana', 'has', 'an', 'elevated', 'risk', 'based', 'on', 'our', 'metrics', '.', 'learn', '.', '.']\n",
            "pred:  0 label:  0 ['photo', 'of', 'several', 'people', 'lying', 'on', 'ground', 'in', 'middle', 'of', 'street', 'allegedly', 'in', 'china', '.']\n",
            "pred:  1 label:  1 ['rt', 'cdc', 'recommends', 'using', 'an', 'household', 'disinfectant', 'to', 'protect', 'against', 'virus', 'that', 'causes', '.', 'visit']\n",
            "pred:  0 label:  0 ['italy', 'registered', '4200', 'new', 'recovering', 'cases', 'from', 'coronavirus', 'in', '24', 'hours', '.']\n",
            "pred:  0 label:  0 ['dr', 'vk', 'srinivas', 'vice', 'president', 'bharat', 'biotech', 'taking', 'coronavirus', 'vaccine', 'clinical', 'trial', 'he', 'is', 'first', 'person', 'in', 'india', 'to', 'take', 'vaccine', 'developed', 'by', 'him', 'and', 'his', 'team', 'in', 'bharat', 'biotech', '.']\n",
            "pred:  1 label:  1 ['there', 'is', 'no', 'one', 'in', 'new', 'zealand', 'receiving', 'care', 'for', '.', 'yesterday', 'our', 'laboratories', 'completed', '1754', 'tests', '.', 'that', 'brings', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to']\n",
            "pred:  0 label:  0 ['for', 'toilet', 'robbing', 'bastards', ',', 'should', 'be', 'ashamed', 'of', 'canons', 'park', 'disgusting', '.']\n",
            "pred:  1 label:  1 ['correct', 'that', 'rt', 'becomes', 'harder', 'to', 'maintain', 'as', 'cases', 'become', 'low', '.', 'currently', 'working', 'on', 'improving', 'our', 'metrics', 'to', 'better', 'measure', 'containment']\n",
            "pred:  1 label:  1 ['there', 'is', 'one', 'remaining', 'significant', 'cluster', 'that', 'is', 'still', 'classified', 'as', 'open', '–', 'st', '.', 'hospital', ';', 'rest', 'home', 'in', 'auckland', '.', 'yesterday', 'our', 'labs', 'completed', 'tests', 'bringing', 'total', 'completed', 'to', 'date', 'to', '.', 'this', 'is', 'much', 'higher', 'than', 'last', 'few', 'weeks', '.']\n",
            "pred:  0 label:  0 ['wuhan', 'locals', 'claim', 'coronavirus', 'has', 'killed', '42', ',', '000', 'people', 'in', 'city', 'alone', ',', 'more', 'than', 'ten', 'times', 'national', 'figure', 'claimed', 'by', 'chinese', 'authorities', '.', '.']\n",
            "pred:  0 label:  0 ['dead', 'bodies', 'wrapped', 'in', 'these', 'white', 'cloths', 'are', 'those', 'of', 'people', 'who', 'died', 'of', 'in', 'osmania', 'hospital', 'in', 'city', 'of', 'hyderabad', 'india', '.']\n",
            "pred:  1 label:  1 ['since', '9th', 'of', 'june', 'we', 'have', 'rolled', 'out', 'testing', 'in', 'managed', 'isolation', 'facilities', 'at', 'around', 'days', '3', 'and', '12', 'of', 'stay', '.', 'some', 'have', 'suggested', 'that', 'everyone', 'should', 'also', 'be', 'tested', 'on', 'arrival', '.', 'please', 'know', 'that', 'everyone', 'who', 'arrives', 'at', 'border', 'goes', 'through', 'a', 'health', 'screen', '.']\n",
            "pred:  1 label:  1 ['per', 'day', 'testing', 'capacity', 'has', 'crossed', '10', 'lakhs', 'tests', '.', 'tests', 'were', 'conducted', 'in', 'last', '24', 'hours', '.', 'weekly', 'average', 'tests', 'have', 'multiplied', 'by', 'more', 'than', '4', 'times', 'since', 'first', 'week', 'of', 'january', '2020', '.', '.']\n",
            "pred:  1 label:  1 ['people', 'who', 'may', 'have', 'coronavirus', 'will', 'be', 'required', 'to', 'by', 'law', '-', 'with', 'those', 'who', 'refuse', 'facing', 'fines', 'of', 'up', 'to', 'in', 'england', '.', 'has', 'more', 'on', 'how', 'new', 'rules', 'will', 'work', '.', 'read', 'more', '.', '.']\n",
            "pred:  0 label:  0 ['videos', 'show', 'clashes', 'arose', 'in', 'paris', 'because', 'new', 'coronavirus', 'was', 'detected', 'in', 'water', '.']\n",
            "pred:  0 label:  0 ['vaccine', 'against', 'new', 'coronavirus', 'has', 'existed', 'since', '2001', '.']\n",
            "pred:  0 label:  0 ['image', 'claims', 'that', 'chroma', 'screen', 'panels', 'are', 'being', 'used', 'to', 'make', 'up', 'coronavirus', 'deaths', '.', '_']\n",
            "pred:  0 label:  0 ['case', 'growth', 'will', 'peak', 'by', 'early', 'may', 'and', 'reduce', 'to', '0', 'by', 'may', '16', 'said', 'v', '.', 'k', '.', 'paul', 'of', 'empowered', 'group', '1', 'on', 'april', '24', '.', '53', 'days', 'since', 'a', 'countrywide', 'cases', 'have', 'risen', 'by', ';', 'for', 'a', 'week', '.']\n",
            "pred:  1 label:  1 ['3', 'stages', 'to', 'brain', 'damage', 'identified', '.', '.']\n",
            "pred:  1 label:  1 ['several', 'dozen', 'of', 'our', 'volunteers', 'spent', 'last', '5', 'days', 'evaluating', 'new', 'data', '.', 'this', 'is', 'white', 'paper', 'we', 'produced', '.', '.', '.']\n",
            "pred:  0 label:  0 ['virus', 'is', 'a', 'blessing', 'that', 'sucks', '.', '”', 'and', 'anonymous']\n",
            "pred:  1 label:  1 ['rt', 'domestic', 'aviation', 'operations', 'soaring', 'nearly', 'passengers', 'took', 'to', 'skies', 'across', 'country', 'on', '6th']\n",
            "pred:  0 label:  0 ['squeezing', 'hot', 'lemon', 'water', 'cures', 'corona', 'virus', 'infection', '.']\n",
            "pred:  0 label:  0 ['meme', 'claims', 'that', 'bill', 'gates', 'was', 'from', 'india', 'for', 'leaving', 'permanently', 'disabled', 'and', 'disabled', 'children', 'a', 'direct', 'result', 'of', 'his', 'polio', 'vaccine', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'putting', 'a', 'national', 'lockdown', ',', 'stay-at-home', 'orders', ',', 'is', 'like', 'house', 'arrest', '.', 'other', 'than', 'slavery', ',', 'which', 'was', 'a', 'different', 'kind', 'of', 'restraint', ',', 'this', 'is', 'greatest', 'intrusion', 'on', 'civil', 'liberties', 'in', 'american', 'history', '.', '\"']\n",
            "pred:  1 label:  1 ['know', 'when', 'to', 'delay', 'your', 'travel', 'to', 'of', '.', 'check', 'these', 'common', 'situations', 'and', 'talk', 'to', 'your', 'healthcare', 'provider', 'if', 'you', 'are', 'unsure', 'whether', 'any', 'of', 'these', 'apply', 'to', 'you', 'or', 'your', 'travel', 'companions', '.', '.', '.']\n",
            "pred:  0 label:  0 ['neighborhood', 'signs', 'asking', 'medical', 'staff', 'not', 'to', 'go', 'home', 'and', 'stay', 'elsewhere', 'in', 'spain', '.']\n",
            "pred:  0 label:  0 ['cocaine', 'can', 'protect', 'against', 'coronavirus']\n",
            "pred:  0 label:  0 ['there', 'are', 'two', 'coronavirus', 'cases', 'in', 'calabria', '.']\n",
            "pred:  0 label:  0 ['cindy', 'mccain', 'appointed', 'to', 'joe', 'transition', 'team', '.', 'no', 'us', 'widow', 'is', 'to', 'identify', 'biden', ',', 'who', 'has', 'deep', 'corrupt', 'ties', 'to', 'that', 'launched', 'on', 'american', 'soil', 'killing', 'thousands', ',', 'others', 'face', 'down', '.', '.']\n",
            "pred:  0 label:  0 ['herd', 'immunity', 'strategy', 'is', 'and', 'unlikely', 'to', 'be', 'successful', 'in', 'uk', ',', 'warn', 'scientists', ',', 'as', 'long', 'sufferers', 'call', 'for', 'more', 'recognition', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'six', 'months', 'before', '?', '?', '?', ',', '?', '?', '?', 'bill', 'gates', 'had', 'negotiated', 'a', 'billion', 'contact', 'tracing', 'deal', 'with', 'democratic', 'congressman', 'sponsor', 'of', 'bill', '.', '?', '?']\n",
            "pred:  0 label:  0 ['if', 'you', 'get', 'coronavirus', 'from', 'chinese', 'food', 'simple', 'cure', 'is', 'to', 'gargle', 'bleach', '.']\n",
            "pred:  0 label:  0 ['man', 'feels', 'eye', 'pain', 'exactly', '45', 'seconds', 'after', 'reading', 'a', 'coronavirus', 'symptom', '.', '.']\n",
            "pred:  1 label:  1 ['our', 'total', 'number', 'of', 'confirmed', 'cases', 'of', 'remains', 'at', '1458', 'which', 'is', 'number', 'we', 'report', 'to', 'world', 'health', 'organization', '.', 'yesterday', 'our', 'laboratories', 'processed', 'tests', 'bringing', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  1 label:  1 ['we', 'have', 'our', 'daily', '4pm', 'update', 'live', 'now', '.', '.', 'tracked', 'people', 'tested', 'total', '.']\n",
            "pred:  1 label:  1 ['has', 'reported', 'recovery', 'of', '684', 'patients', 'managed', 'in', 'community', 'in', 'line', 'with', 'new', 'case', 'management', 'guidelines', '.', 'a', 'breakdown', 'of', 'cases', 'by', 'state', 'can', 'be', 'found', 'via', '.', '.']\n",
            "pred:  0 label:  0 ['doctor', 'operated', 'and', 'took', 'out', 'kidney', 'of', 'a', 'patient']\n",
            "pred:  1 label:  1 ['second', 'round', 'of', 'countrywide', 'led', 'by', 'icmr', 'has', 'been', 'successfully', 'completed', '.', 'final', 'phase', 'analysis', 'of', 'survey', 'is', 'now', 'underway', 'and', 'will', 'offer', 'a', 'comparison', 'with', 'results', 'of', 'first', 'survey']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'future', 'vaccine', 'will', 'come', 'with', 'id', '?', '?', '\"', 'a', 'mark', '.', 'id', 'tells', 'everyone', 'you', 'are', 'free', 'of', '.', '?', '?']\n",
            "pred:  0 label:  0 ['suggests', 'trump', 'urged', 'sick', 'people', 'to', 'get', 'out', 'and', 'vote', 'during', 'pandemic']\n",
            "pred:  0 label:  0 ['herman', 'cain', ',', 'one-time', 'republican', 'presidential', 'candidate', ',', 'died', 'of', 'colon', 'cancer', ',', 'not', '.']\n",
            "pred:  1 label:  1 ['drew', 'a', 'difference', 'between', 'these', 'protests', 'against', 'police', 'brutality', 'and', 'protests', 'earlier', 'this', 'spring', 'which', 'opposed', 'mask', 'mandates', 'and', 'rules', '.', '”']\n",
            "pred:  1 label:  1 ['today', 'there', 'are', '10', 'people', 'in', 'hospital', 'who', 'have', '.', 'three', 'people', 'are', 'in', 'auckland', 'city', 'hospital', 'four', 'people', 'in', 'middlemore', 'two', 'people', 'in', 'north', 'shore', 'hospital', 'and', 'one', 'person', 'in', 'waikato', 'hospital', '.', 'he', 'new', 'person', 'in', 'auckland', 'city', 'hospital', 'is', 'linked', 'to', 'community', 'cluster']\n",
            "pred:  0 label:  0 ['scientists', 'have', 'developed', 'a', 'cure', 'for', 'coronavirus', '.', '_']\n",
            "pred:  0 label:  0 ['children', 'can', 'be', 'taken', 'away', 'for', 'test', 'without', 'parental', 'consent', 'in', 'uk', 'is', 'fake']\n",
            "pred:  0 label:  0 ['how', 'much', 'you', 'wanna', 'bet', 'this', 'is', 'be', 'cure', 'for', 'coronavirus', 'disease', '?', 'all', 'i', 'ask', 'is', 'to', 'make', 'sure', 'i', 'get', 'credit', 'for', 'coming', 'up', 'with', 'this', 'magical', 'cure', 'all']\n",
            "pred:  0 label:  0 ['an', 'article', 'published', 'on', 'a', 'website', 'called', '.', 'co', '.', 'in', 'is', 'claiming', 'prime', 'minister', 'narendra', 'modi', 'is', 'likely', 'to', 'announce', 'emergency', 'in', 'india', 'under', 'article', '360', '—', 'financial', 'emergency', '.']\n",
            "pred:  0 label:  0 ['manchester', 'united', 'fans', 'wondering', 'if', 'a', 'global', 'coronavirus', 'pandemic', 'could', 'prevent', 'liverpool', 'winning', 'league', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'staying', 'home', 'is', 'best', 'way', 'to', 'slow', 'spread', 'of', '.', 'it', 'can', 'be', 'hard', 'to', 'remain', 'apart', 'from', 'loved', 'ones', 'during']\n",
            "pred:  0 label:  0 ['according', 'to', 'civil', 'registry', 'record', 'number', 'of', 'respiratory', 'syndrome', 'deaths', 'decreased', 'and', 'not', 'increased', 'between', '2019', 'and', '2020', 'in', 'ceará', 'brazil', '.', 'time', 'period', 'considered', 'was', 'march', '16', 'to', 'may', '10', '.']\n",
            "pred:  1 label:  1 ['a', 'few', 'thoughts', 'on', 'why', 'we', 'counted', 'more', 'than', '6600', 'new', 'tests', '1', '.', 'wa', 'which', 'accounts', 'for', 'nearly', 'of', 'all', 'reported', 'tests', 'in', 'us', 'did', 'not', 'report', 'before', '4', 'pm', 'et', 'yesterday', '.', '2', '.', 'ny', 'reported', 'total', 'tests', 'last', 'night', 'for', 'first', 'time', '.']\n",
            "pred:  1 label:  1 ['some', 'states', 'reactivating', 'emergency', 'plans', 'for', 'crisis', 'care', 'and', 'delaying', 'elective', 'surgeries', '.', '.']\n",
            "pred:  1 label:  1 ['as', 'of', '8', 'aug', '2020', 'has', 'tested', 'persons', 'for', '.']\n",
            "pred:  1 label:  1 ['achieves', 'a', 'record', 'number', 'of', 'tests', 'in', 'a', 'single', 'day', 'on', 'september', '19', '2020', '.', '.']\n",
            "pred:  0 label:  0 ['chinese', 'president', 'xi', 'jinping', 'urging', 'africans', 'to', 'reject', 'a', 'coronavirus', 'vaccine', 'that', 'has', 'killed', 'all', 'animals', 'it', 'has', 'been', 'tested', '.']\n",
            "pred:  1 label:  1 ['california', 'is', 'now', 'in', 'united', 'states', 'for', 'total', 'deaths', 'behind', 'new', 'york', 'and', 'new', 'jersey', '.', 'we', 'hope', 'that', 'declining', 'case', 'counts', 'in', 'ca', 'will', 'be', 'reflected', 'in', 'falling', 'deaths', 'within', 'next', '2-3', 'weeks', '.', '.']\n",
            "pred:  0 label:  0 ['cdc', 'said', 'it', '?', '?', '?', 'made', 'a', 'mistake', '?', '?', 'and', 'reduced', 'its', 'count', 'of', 'florida', 'cases', 'from', '90', ',', '000', 'to', '11', ',', '000', '.']\n",
            "pred:  1 label:  1 ['by', 'conducting', 'close', 'to', '5', 'crore', 'tests', '.', '1', '.', '33', 'crore', 'tests', 'were', 'conducted', 'in', 'last', 'week', 'alone', '.', '.', '.']\n",
            "pred:  0 label:  0 ['trump', 'announces', 'a', 'cure', 'for', '.']\n",
            "pred:  0 label:  0 ['video', 'of', 'migrants', 'in', 'a', 'overcrowded', 'train', 'from', 'mumbai', 'to', 'west', 'bengal', 'on', '10', 'may', '2020', 'amid', 'social', 'distancing', '.']\n",
            "pred:  1 label:  1 ['recovery', 'rate', 'improves', 'to', '78', '.', 'as', 'on', 'september', '17', '2020', 'improvement', 'in', 'recovery', 'rate', 'since', 'initiation', 'on', 'march', '25', '2020', 'via', '.']\n",
            "pred:  1 label:  1 ['those', 'feel', 'most', 'solid', '.', 'still', 'not', 'sure', 'how', 'much', 'coverage', 'each', 'state', 'is', 'getting', 'of', 'its', 'hospitals', '.']\n",
            "pred:  1 label:  1 ['from', 'tomorrow', 'we', 'are', 'all', 'being', 'asked', 'not', 'visit', 'each', 'homes', '.', 'first', 'minister', 'nicola', 'sturgeon', 'explains', 'new', 'restriction', 'banning', 'people', 'from', 'meeting', 'other', 'households', 'indoors', '.', 'get', 'latest', 'on', 'new', '.', '.']\n",
            "pred:  1 label:  1 ['4', 'were', 'granted', 'day', 'visits', 'to', 'see', 'a', 'family', 'member', 'who', 'was', 'closing', 'to', 'dying', '.', 'they', 'returned', 'to', 'facility', 'at', 'end', 'of', 'each', 'day', '.', 'one', 'person', 'was', 'granted', 'an', 'exemption', 'because', 'of', 'a', 'terminal', 'medical', 'condition', '.', 'these', 'people', 'were', 'granted', 'an', 'exemption', 'for', 'exceptional', 'circumstances', '.']\n",
            "pred:  1 label:  1 ['one', 'major', 'caveat', 'texas', 'reported', 'a', 'backlog', 'of', '5k', 'cases', 'which', 'are', 'included', 'in', 'numbers', '.', 'even', 'for', 'a', 'monday', 'which', 'is', 'subject', 'to', 'weekend', 'lag', 'this', 'is', 'a', 'very', 'very', 'small', 'number', 'of', 'cases', '.']\n",
            "pred:  0 label:  0 ['i', 'thought', 'corona', 'affects', 'only', 'humans', '.', '.']\n",
            "pred:  1 label:  1 ['575', 'new', 'cases', 'of', ';', 'confirmed', 'discharged', '709', 'deaths', '.']\n",
            "pred:  1 label:  1 ['rt', 'protect', 'yourself', 'and', 'others', 'from', 'getting', 'wash', 'your', 'hands', 'after', 'coughing', 'or', 'sneezing', 'when', 'caring', 'for']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '5', '.', '8', 'million', 'tests', 'up', '200k', 'from', 'yesterday', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'for', 'details', '.']\n",
            "pred:  0 label:  0 ['moms', 'form', 'vigilante', 'gang', 'to', 'make', 'sure', 'you', 'get', 'that', 'cough', 'checked', 'out', '.', '.']\n",
            "pred:  1 label:  1 ['is', 'shocked', 'that', 'pandemic', 'response', 'is', 'worst', 'in', 'world', '.', '.', 'via']\n",
            "pred:  1 label:  1 ['ramped', 'up', 'hospital', 'infrastructure', ';', 'emphasis', 'on', 'efficient', 'treatment', 'of', 'hospitalized', 'patients', 'through', 'standard', 'of', 'care', 'protocol', 'advised', 'by', 'centre', 'have', 'effectively', 'ensured', 'improvement', 'in', 'recovery', 'rate', '.', '.']\n",
            "pred:  0 label:  0 ['muslims', 'are', 'spreading', 'coronavirus', 'by', 'leaving', 'money', 'strewn', 'on', 'streets', '.']\n",
            "pred:  1 label:  1 ['new', 'report', 'is', 'first', 'to', 'look', 'at', 'illness', 'in', 'kids', 'in', 'u', '.', 's', '.', 'among', 'more', 'than', '2500', 'cases', 'of', 'in', 'children', 'in', 'u', '.', 's', '.', 'three', 'children', 'reportedly', 'died', '.', '.', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '15', '.', '2', 'million', 'tests', 'up', 'just', 'from', 'yesterday', 'way', 'below', 'average', '.', 'not', 'clear', 'why', 'tests', 'have', 'fallen', 'so', 'far', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'for', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'find', 'important', 'info', '–', 'from', 'social', 'distancing', 'to', 'effective', 'cleaning', 'to', 'maintaining', 'your', 'mental', 'health', '–', 'in', 'our']\n",
            "pred:  0 label:  0 ['harr', 'roz', 'corona', 'cases', 'h', ',', 'delhi', 'mai', '.', 'cm', 'sir', 'said', '\"', 'condition', 'are', 'under', 'control', 'and', 'we', 'are', 'in', 'best', 'lockdown', '4', '.', 'in', 'today', ',', 'delhi', 'had', 'open', 'park', ',', 'bazar', 'without', 'social', 'distancing', 'and', 'also', 'we', 'know', 'ito', 'jam', ',', 'border', 'jam', '.', 'corona', 'cases']\n",
            "pred:  0 label:  0 ['mass', 'disinfection', 'of', 'people', 'using', 'a', 'chemical', 'solution', 'will', 'eradicate', '.']\n",
            "pred:  0 label:  0 ['gargling', 'by', 'salt', 'water', 'and', 'inhaling', 'hot', 'water', 'cures', '.']\n",
            "pred:  0 label:  0 ['who', 'has', 'declared', 'an', 'endemic', 'disease', '.']\n",
            "pred:  0 label:  0 ['president', 'trump', 'presented', 'with', 'panda', 'from', 'xi', 'jinping', '.']\n",
            "pred:  0 label:  0 ['bill', 'gates', 'claims', 'and', 'announces', 'that', 'vaccinations', 'will', 'be', 'mandatory', 'for', 'all', '.']\n",
            "pred:  1 label:  1 ['highlights', 'of', 'new', 'nationwide', 'measures', 'are', 'as', 'follows', ';', 'there', 'will', 'be', 'an', 'overnight', 'curfew', 'from', '8pm', 'to', '6am', '.', 'this', 'means', 'all', 'movement', 'will', 'be', 'prohibited', 'during', 'this', 'period', 'except', 'essential', 'services', ';']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'cures', 'this', '?', '?', '?', 'virus', '.', '?', '?', '?', 'it', 'just', 'so', 'happens', 'this', 'is', 'treatment', 'used', 'for', 'radiation', '?', '?']\n",
            "pred:  0 label:  0 ['10', 'million', 'people', 'contracted', 'tuberculosis', 'last', 'year', '.', '1', '.', '5', 'million', 'people', 'died', '.', 'did', 'you', 'even', 'know', '?', 'were', 'you', 'scared', 'for', 'your', 'life', '?', 'did', 'we', 'wear', 'masks', ',', 'close', 'economy', ',', 'cancel', 'schools', ',', 'and', 'ruin', 'small', 'businesses', '?', 'no', '.', 'why', '?', 'because', 'media', 'tell', 'you', 'to', 'be']\n",
            "pred:  1 label:  1 ['daily', 'mortality', 'growth', 'rate', 'declined', 'by', '~', '6', '.', '1', '%', 'points', 'and', 'hospitalization', 'growth', 'rate', 'declined', 'by', '~', '8', '.', '4', '%', 'points', 'in', 'states', 'with', 'stay', 'at', 'home', 'orders', '.', '.', '.']\n",
            "pred:  1 label:  1 ['an', 'update', 'on', 'in', 'terms', 'of', 'compassionate', 'exemptions', 'we', 'are', 'waiting', 'on', '2', 'people', 'to', 'return', 'tests', 'and', 'we', 'continue', 'to', 'work', 'with', 'enforcement', 'services', 'to', 'follow', 'up', 'on', 'one', 'person', '.', '40', 'have', 'returned', 'negative', 'results', 'and', '11', 'are', 'not', 'being', 'tested', 'for', 'a', 'range', 'of', 'reasons', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', '.', 's', '.', 'rep', '.', 'conor', 'said', 'he', 'would', 'not', 'vote', 'for', 'for', 'speaker', ',', 'and', 'did', '.', '?', '?']\n",
            "pred:  0 label:  0 ['sweden', 'to', 'be', 'extradited', 'to', 'united', 'states', 'for', 'violating', 'american', 'guidelines', '.']\n",
            "pred:  1 label:  1 ['of', 'new', 'recovered', 'cases', 'are', 'being', 'reported', 'from', 'ten', '.', 'maharashtra', 'continues', 'to', 'lead', 'with', 'more', 'than', '.', 'new', 'recovered', 'patients', '.', 'andhra', 'pradesh', 'contributed', 'more', 'than', '10000', 'to', 'single', 'day', 'recoveries', '.', '.']\n",
            "pred:  1 label:  1 ['97', 'new', 'cases', 'of', ';', 'confirmed', 'discharged', '1098', 'deaths', '.']\n",
            "pred:  1 label:  1 ['clothing', 'is', 'unlikely', 'to', 'be', 'a', 'huge', 'source', 'of', 'transmission', '.', '.']\n",
            "pred:  0 label:  0 ['in', 'kanpur', 'india', 'left', 'outside', 'hopsital', 'without', 'any', 'treatment', '.']\n",
            "pred:  1 label:  1 ['john', 'brooks', 'is', 'chief', 'medical', 'officer', 'for', 'response', '.', 'he', ';', 'his', 'team', 'sift', 'through', 'newly', 'published', 'reports', 'on', 'in', 'hopes', 'of', 'applying', 'findings', 'to', 'slow', 'spread', 'of', '.', 'read', '.', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'update', 'total', 'of', '1172', 'deaths', 'have', 'been', 'registered', 'in', 'past', '24', 'hours', 'total', 'deaths', 'are']\n",
            "pred:  1 label:  1 ['there', 'are', '82', 'people', 'in', 'community', 'cluster', 'who', 'remain', 'at', 'auckland', 'quarantine', 'facility', 'that', 'includes', '59', 'people', 'who', 'have', 'tested', 'positive', 'for', 'and', 'their', 'household', 'contacts', '.', 'that', 'number', 'continues', 'to', 'decline', 'as', 'cases', 'recover', 'and', 'reach', 'end', 'of', 'their', 'isolation', 'period']\n",
            "pred:  1 label:  1 ['labour', 'leader', 'sir', 'keir', 'starmer', 'says', 'he', 'that', 'head', 'of', 'test', 'and', 'trace', 'dido', 'harding', 'and', 'her', 'team', 'were', 'unaware', 'a', 'second', 'spike', 'would', 'occur', '.', 'latest', 'on', 'response', '.', '.']\n",
            "pred:  1 label:  1 ['a', 'second', 'national', 'lockdown', 'could', 'be', 'needed', 'if', 'latest', 'restrictions', 'do', 'not', 'work', 'foreign', 'secretary', 'dominic', 'raab', 'has', 'told', 'sky', 'news', '.']\n",
            "pred:  1 label:  1 ['health', 'ministry', 'urges', 'to', 'mandatorily', 'retest', 'all', 'symptomatic', 'negative', 'cases', 'of', 'rapid', 'antigen', 'tests', 'through', 'rt-pcr', '.', '.']\n",
            "pred:  1 label:  1 ['apart', 'from', 'ones', 'who', 'had', 'a', 'compassionate', 'exemption', 'or', 'who', 'have', 'gone', 'off', 'shore', 'everyone', 'in', 'this', 'group', 'completed', '14', 'day', 'isolation', '.', 'however', 'we', 'are', 'following', 'them', 'up', 'and', 'testing', 'them', 'to', 'close', 'loop', 'and', 'strengthen', 'our', 'system', '.', 'these', 'numbers', 'will', 'be', 'updated', 'daily', '.']\n",
            "pred:  0 label:  0 ['says', 'masks', 'to', 'reduce', 'transmission', 'violate', 'osha', 'oxygen', 'levels', '.', '”']\n",
            "pred:  0 label:  0 ['president', 'donald', 'trump', 'got', 'a', 'moroccan', 'sheikh', 'to', 'put', 'a', 'spell', 'on', 'him', 'that', 'prevents', '.']\n",
            "pred:  1 label:  1 ['nhs', 'services', 'are', 'being', 'put', 'at', 'risk', 'because', 'a', 'lack', 'of', 'coronavirus', 'testing', 'is', 'forcing', 'many', 'staff', 'to', 'when', 'they', 'could', 'be', 'at', 'work', '.']\n",
            "pred:  0 label:  0 ['rt', 'we', 'president', 'oshkosh', 'speech', 'digging', 'into', 'claims', 'on', 'numbers', 'farmers']\n",
            "pred:  0 label:  0 ['who', 'declared', 'is', 'abbreviation', 'for', 'china', 'outbreak', 'virus', 'in', 'december', '2019', '.']\n",
            "pred:  1 label:  1 ['rt', 'waiting', 'for', 'a', 'test', 'result', 'and', 'wondering', 'what', 'you', 'should', 'do', 'next', '?', 'take', '3', 'key', 'stay', 'home', 'think', 'about']\n",
            "pred:  1 label:  1 ['as', 'at', '11th', 'april', 'there', 'are', '318', 'confirmed', 'cases', '70', 'discharged', '10', 'deaths', '174', '56', '20', '12', '11', '7', '6', '6', 'akwa', '5', '3', '2', '2', '2', '2', '1', '1', '1']\n",
            "pred:  1 label:  1 ['cases', 'per', 'million', 'population', 'is', 'amongst', 'lowest', 'in', 'world', ';', 'which', 'stands', 'at', '3102', 'for', 'india', 'and', 'for', 'secretary', '.']\n",
            "pred:  0 label:  0 ['multiple', 'facebook', 'and', 'twitter', 'posts', 'shared', 'thousands', 'of', 'times', 'by', 'australian', 'social', 'media', 'users', 'claim', 'people', 'with', 'pre-existing', 'respiratory', 'conditions', 'will', 'be', 'given', 'a', 'of', 'medication', 'from', 'their', 'general', 'practitioners', 'during', 'novel', 'coronavirus', 'pandemic', '.']\n",
            "pred:  0 label:  0 ['herd', 'immunity', 'strategy', 'is', 'and', 'unlikely', 'to', 'be', 'successful', 'in', 'uk', ',', 'warn', 'scientists', ',', 'as', 'long', 'sufferers', 'call', 'for', 'more', 'recognition', '.']\n",
            "pred:  1 label:  1 ['on', 'friday', 'there', 'were', '133', 'people', 'swabbed', 'over', 'a', 'two-hour', 'period', 'at', 'new', 'world', 'in', 'new', 'plymouth', '.', 'testing', 'centre', 'in', 'manurewa', 'had', '153', 'people', 'swabbed', 'yesterday', 'and', '326', 'people', 'were', 'tested', 'yesterday', 'at', 'pop-up', 'clinic', 'in', 'christchurch', '.']\n",
            "pred:  1 label:  1 ['number', 'of', 'recovered', 'cases', 'remains', 'at', '1482', '.', 'there', 'are', 'no', 'additional', 'deaths', 'to', 'report', '.', 'there', 'is', 'no', 'one', 'in', 'new', 'zealand', 'receiving', 'care', 'for', '.', 'yesterday', 'our', 'laboratories', 'completed', 'tests', 'bringing', 'total', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  1 label:  1 ['nation', 'passed', '10000', 'deaths', 'today', '.', 'new', 'york', 'stands', 'at', 'deaths', '.', 'new', 'jersey', 'passed', '1000', '.', '16', 'states', 'have', 'lost', 'more', 'than', '100', 'people', 'to', '.']\n",
            "pred:  1 label:  1 ['new', 'local', 'restrictions', 'are', 'being', 'introduced', 'in', 'northeast', 'england', '-', 'including', 'curfews', 'for', 'bars', 'and', 'pubs', 'and', 'a', 'ban', 'on', 'people', 'mixing', 'with', 'others', 'outside', 'their', 'household', '.']\n",
            "pred:  1 label:  1 ['zooming', 'back', 'out', '15', 'states', 'have', 'now', 'set', 'their', 'record', 'for', 'reported', 'cases', 'since', 'june', '19', '.', 'all', 'but', 'one', 'is', 'in', 'south', 'and', 'west', '.', 'we', 'know', 'some', 'people', 'do', 'classify', 'mo', 'in', 'south', '.', ')', '.']\n",
            "pred:  0 label:  0 ['in', 'india', 'congress', 'approval', 'is', 'required', 'to', 'use', 'funds', 'from', '.']\n",
            "pred:  0 label:  0 ['flabby', 'coronavirus', 'looking', 'forward', 'to', 'getting', 'back', 'to', 'gym', '.', '.']\n",
            "pred:  1 label:  1 ['how', 'did', 'alaska', 'test', 'more', 'per', 'capita', 'than', 'any', 'other', 'state', ';', 'have', 'lowest', 'deaths', 'per', 'capita', '?', 'state', 'contracted', 'with', 'a', 'local', 'manufacturer', 'to', '3d', 'print', 'testing', 'swabs', 'and', 'deploy', 'rapid', 'tests', 'that', 'they', 'processed', 'in-state', '.', '.']\n",
            "pred:  1 label:  1 ['vte', 'risk', 'is', '~', 'lower', 'in', 'patients', 'with', 'ra', 'taking', 'tnf', 'inhibitors', 'in', 'new', 'research', '.', '.']\n",
            "pred:  0 label:  0 ['these', 'pictures', 'are', 'of', 'dr', 'v', '.', 'k', 'srinivas', 'of', 'bharat', 'biotech', 'taking', 'first', '&', 'second', 'doses', 'of', 'corona', 'vaccine', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'people', 'are', 'going', 'door', 'to', 'door', 'in', 'colorado', 'springs', 'stating', 'they', 'are', 'testing', 'so', 'they', 'can', 'rob', '?', '?']\n",
            "pred:  0 label:  0 ['another', 'coronavirus', 'figure', 'on', 'time', '100', 'list', '.', 'shi', ',', 'aka', 'woman', ',', '\"', 'of', 'wuhan', 'institute', 'of', 'virology', '.', '.']\n",
            "pred:  1 label:  1 ['sadly', 'cases', 'are', 'rising', 'in', 'many', 'states', 'beyond', 'florida', 'texas', 'and', 'arizona', '.', '23', 'states', 'have', 'seen', 'their', 'highest', 'day', 'of', 'cases', 'in', 'past', '2', 'weeks', '.', '.']\n",
            "pred:  1 label:  1 ['there', 'is', 'no', 'one', 'in', 'new', 'zealand', 'receiving', 'care', 'for', '.', 'yesterday', 'our', 'laboratories', 'completed', '1620', 'tests', 'which', 'brings', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to']\n",
            "pred:  0 label:  0 ['a', 'quote', 'by', 'doctor', 'alberto', 'zangrillo', 'saying', 'that', 'is', 'a', 'consequence', 'of', 'immigration', '.']\n",
            "pred:  0 label:  0 ['president', 'trump', 'has', 'said', 'that', 'of', 'all', 'cases', 'are', 'totally', 'harmless', '.', 'a', 'lot', 'more', 'to', 'be', 'learned', 'about', 'coronavirus', 'but', 'nothing', 'says', 'that', 'harmless', 'for', 'of', 'people', 'it', 'touches', '.', '.', '.']\n",
            "pred:  0 label:  0 ['bill', 'gates', 'already', 'has', 'his', 'vaccine', 'ready', 'for', 'you', 'ed', '.', ')', 'and', 'patent', '.']\n",
            "pred:  0 label:  0 ['developed', 'a', 'coronavirus', 'vaccine', '.', '_']\n",
            "pred:  0 label:  0 ['omg', '.', '.', '.', 'are', 'you', 'frigging', 'kidding', 'me', '.', '.', '.', 'just', 'like', 'trump', 'knowing', 'how', 'deadly', 'coronavirus', 'was', '.', '.', '.', '.', 'same', 'goes', 'for', 'police', 'that', 'night', '.', '.', '.', 'they', 'said', 'knocked', '.', '.', '.', 'it', 'was', 'their', 'words', 'against', 'her', 'distraught', 'boyfriend']\n",
            "pred:  0 label:  0 ['exposure', 'notification', 'from', 'google', '&', 'apple', 'is', 'a', 'sensor', 'inserted', 'to', 'trace', 'every', 'phone', '.']\n",
            "pred:  0 label:  0 ['man', 'considering', 'going', 'into', 'marble', 'production', '.']\n",
            "pred:  0 label:  0 ['indian', 'home', 'minister', 'amit', 'shah', 'said', 'that', 'government', 'transferred', 'financial', 'aid', 'of', 'rs', '530', 'million', 'into', 'bank', 'accounts', 'of', '410', 'million', 'people', 'during', 'coronavirus', 'pandemic', 'which', 'means', 'rs', '1', '.', '29', 'for', 'each', 'person', '.']\n",
            "pred:  0 label:  0 ['you', 'will', 'get', 'with', 'if', 'you', 'eat', 'sugar', 'drink', 'cold', 'water', 'or', 'take', 'a', 'shower', 'at', 'night']\n",
            "pred:  0 label:  0 ['millions', 'of', 'app', 'send', 'in', '3900', 'of', '8', 'different', 'type', 'of', 'rashes', ',', 'so', 'now', 'a', 'symptom', '.']\n",
            "pred:  1 label:  1 ['washington', 'has', 'acknowledged', 'their', 'issues', 'with', 'reporting', 'negative', 'tests', '.', 'hopefully', 'we', 'see', 'those', 'numbers', 'return', 'to', 'normal', 'in', 'near', 'future', '.', '.']\n",
            "pred:  0 label:  0 ['united', 'states', 'of', 'america', 'threatened', 'iraq', 'with', 'coronavirus', 'in', '1990', '.']\n",
            "pred:  0 label:  0 ['high', 'temperatures', 'kill', 'novel', 'corona', 'virus', '.']\n",
            "pred:  1 label:  1 ['239', 'new', 'cases', 'of', ';', 'cases', 'of', 'in', 'nigeria', '745', '128', '.']\n",
            "pred:  0 label:  0 ['lockdown', 'is', 'only', 'for', 'hindus', '.']\n",
            "pred:  0 label:  0 ['video', 'of', 'people', 'in', 'italy', 'offering', 'namaz', 'on', 'streets', 'due', 'to', '.']\n",
            "pred:  0 label:  0 ['france', 'forces', 'patients', 'of', 'dual', 'nationality', 'to', 'leave', 'its', 'hospitals', 'to', 'be', 'able', 'to', 'treat', 'french', 'during', 'outbreak', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'of', 'dead', 'bodies', 'in', 'black', 'bags', 'allegedly', 'from', 'hospitals', 'in', 'madrid', 'barcelona', 'or', 'new', 'york', '.']\n",
            "pred:  1 label:  1 ['and', 'an', 'important', 'story', 'is', 'very', 'different', 'in', 'different', 'regions', 'of', 'country', '.', 'outside', 'cases', 'are', 'not', 'really', 'declining', '.', '.']\n",
            "pred:  0 label:  0 ['pandemic', 'was', 'planned', 'and', 'a', 'preparedness', 'exercise', 'called', 'event', '201', 'predicted', 'it', '.']\n",
            "pred:  0 label:  0 ['a', 'post', 'says', 'that', 'on', '793', 'people', 'died', 'in', 'italy', 'and', 'included', '232', 'kids', '.']\n",
            "pred:  0 label:  0 ['democrats', 'are', 'pushing', 'for', 'an', 'implanted', 'microchip', 'in', 'humans', 'and', 'everyone', 'to', 'be', 'vaccinated', '.', '”']\n",
            "pred:  0 label:  0 ['corona', 'era', ',', 'india', 'surpassed', 'other', 'countries', 'in', 'terms', 'of', 'gdp', '.']\n",
            "pred:  0 label:  0 ['china', 'tv', ':', 'us', 'pushed', 'out', 'vaccine', 'so', 'quickly', ',', 'that', 'only', 'means', 'they', 'have', 'been', 'working', 'on', 'it', 'way', 'before', 'pandemic', '.', '”', 'we', 'can', 'conclude', 'that', 'us', 'had', 'this', 'virus', 'in', 'their', 'possession', 'long', 'ago', '.']\n",
            "pred:  0 label:  0 ['president', 'trump', 'says', 'that', 'he', 'does', 'not', 'take', 'any', 'responsibility', 'for', '2nd', 'wave', 'of', 'coronavirus', 'pandemic', '.']\n",
            "pred:  0 label:  0 ['new', 'corona', 'severe', 'hunger', 'for', 'bats', '?', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'v', '.', 'has', 'virus', 'met', 'its', 'match', '?', '.']\n",
            "pred:  0 label:  0 ['models', 'projecting', 'deaths', '?', '?', '?', 'are', 'talking', 'about', 'without', 'mitigation', '.', '?', '?']\n",
            "pred:  1 label:  1 ['rt', 'total', 'confirmed', 'cases', 'cases', 'samples', 'tested']\n",
            "pred:  0 label:  0 ['who', 'approved', 'home', 'remedy', 'made', 'with', 'pepper', 'ginger', 'juice', 'and', 'honey', 'as', 'a', 'cure', 'for', '.']\n",
            "pred:  0 label:  0 ['this', 'virus', 'must', 'be', 'old', 'because', 'label', 'on', 'a', 'bottle', 'of', 'disinfectant', 'from', '2015', 'shows', '.']\n",
            "pred:  0 label:  0 ['irish', 'lads', 'on', 'quarantined', 'coronavirus', 'cruise', 'having', 'mad', '.']\n",
            "pred:  1 label:  1 ['healthcare', 'cdc', 'offers', 'trainings', 'on', 'clinical', 'care', 'and', 'infection', 'control', 'ppe', 'and', 'emergency', 'preparedness', 'and', 'response', 'specific', 'to', '?', 'find', 'recently', 'recorded', 'webinars', 'and', 'online', '.', '.', '.']\n",
            "pred:  0 label:  0 ['former', 'colombian', 'president', 'and', 'current', 'senator', 'álvaro', 'uribe', 'tested', 'positive', 'for']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'msnbc', 'actually', 'said', 'on', 'air', '?', '?', '?', 'i', 'hope', 'enough', 'people', 'die', 'from', 'coronavirus', 'that', 'it', 'harms', 'trump', '?', '?', '?', 's', 're-election', '.', '?', '?', '?', '?', '?']\n",
            "pred:  0 label:  0 ['outbreak', 'of', 'idiocy', 'spreading', '10', ',', '000', 'times', 'faster', 'than', 'coronavirus', '.', '.']\n",
            "pred:  0 label:  0 ['trump', 'claims', 'speaker', 'nancy', 'pelosi', 'deleted', 'a', 'tweet', 'about', 'her', 'visit', 'to', 'chinatown', 'in', 'february']\n",
            "pred:  1 label:  1 ['on', 'thursday', '21', 'may', 'bars', 'will', 'be', 'able', 'to', 'open', 'with', 'all', 'requirements', 'set', 'out', 'last', 'week', '.', 'we', 'have', 'left', 'bars', 'until', 'last', 'as', 'they', 'pose', 'most', 'risk', '.', 'we', 'need', 'to', 'limit', 'our', 'social', 'gatherings', 'for', 'now', '.', 'parties', 'or', 'big', 'celebrations', 'will', 'need', 'to', 'be', 'limited', 'to', '10', 'people', 'for', 'now', '.']\n",
            "pred:  1 label:  1 ['rhondda', 'cynon', 'taff', 'to', 'go', 'into', 'local', 'lockdown', 'amid', 'rise', 'in', 'cases', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'crisis', 'in', 'italy', 'leads', 'donald', 'trump', 'to', 'close', 'all', 'us', 'pizzerias', '.', '.']\n",
            "pred:  0 label:  0 ['patient', 'zero', 'had', 'sex', 'with', 'bats', '.', '.', '.']\n",
            "pred:  0 label:  0 ['local', 'man', 'appoints', 'himself', 'foremost', 'expert', 'on', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['cdc', 'does', 'not', 'recommend', 'wearing', 'masks', '.']\n",
            "pred:  0 label:  0 ['is', 'being', 'listed', 'as', 'cause', 'of', 'death', 'for', 'people', 'who', 'did', 'not', 'have', 'illness', 'when', 'they', 'died', 'in', 'ireland', '.']\n",
            "pred:  1 label:  1 ['since', '16', 'june', 'when', 'we', 'had', 'two', 'cases', 'returned', 'from', 'uk', 'had', 'more', 'than', '45000', 'tests', 'completed', 'in', 'nz', '.', 'only', 'cases', 'we', 'have', 'detected', 'are', 'those', 'in', 'managed', 'isolation', 'facilities', '.', 'those', 'two', 'cases', 'did', 'draw', 'our', 'attention', 'to', 'a', 'gap', 'in', 'our', 'systems', '–', 'we', 'moved', 'quickly', 'to', 'remedy', 'that']\n",
            "pred:  1 label:  1 ['cases', 'in', 'india', 'on', 'september', '9', '.', '.', 'active', 'cases', '.', 'deaths', 'total', 'confirmed', 'cases', '=', '.']\n",
            "pred:  1 label:  1 ['recovery', 'rate', 'improves', 'to', '80', '.', 'as', 'on', 'september', '21', '2020', 'improvement', 'in', 'recovery', 'rate', 'since', 'initiation', 'on', 'march', '25', '2020', 'via', '.']\n",
            "pred:  1 label:  1 ['this', 'chart', 'makes', 'clear', 'what', 'has', 'been', 'true', 'all', 'different', 'parts', 'of', 'country', 'have', 'experienced', 'this', 'pandemic', 'in', 'different', 'ways', 'at', 'different', 'times', '.', 'after', 'experiencing', 'relatively', 'fewer', 'deaths', 'in', 'spring', 'south', 'has', 'entered', 'a', 'new', 'phase', '.', '.']\n",
            "pred:  0 label:  0 ['tens', 'of', 'thousands', 'of', 'people', 'are', 'moving', 'to', 'laos', '.']\n",
            "pred:  0 label:  0 ['experts', 'in', 'china', 'found', 'that', 'coronavirus', 'can', 'be', 'spread', 'through', 'houseflies', '.']\n",
            "pred:  1 label:  1 ['lag', 'cases', 'on', 'way', 'up', 'and', 'down', 'these', 'still', 'rising', 'hospitalization', 'data', 'remains', 'incomplete', 'and', 'unreliable', 'and', 'added', 'a', 'few', 'more', 'states', 'to', 'our', 'watch', 'list', '.', 'post', 'has', 'all', 'details', '.']\n",
            "pred:  0 label:  0 ['deaths', 'in', 'brazilian', 'state', 'of', 'são', 'paulo', 'fell', 'after', 'audits', 'conducted', 'by', 'new', 'minister', 'of', 'health', '.']\n",
            "pred:  1 label:  1 ['three', 'cases', 'are', 'a', 'man', 'in', 'his', '30s', 'and', 'two', 'children', 'who', 'arrived', 'together', 'from', 'dubai', 'on', 'september', '9', '.', 'they', 'have', 'been', 'in', 'managed', 'isolation', 'at', 'ibis', 'hotel', 'in', 'rotorua', 'and', 'tested', 'positive', 'for', 'during', 'routine', 'testing', 'around', 'day', '3', 'of', 'their', 'stay', '.']\n",
            "pred:  1 label:  1 ['over', '11', '.', '7', 'lakh', 'tests', 'were', 'done', 'in', 'last', '24', 'hours', '.', 'with', 'this', 'achievement', 'cumulative', 'tests', 'are', 'more', 'than', '4', '.', '5', 'crore', '.', 'details', ':', '.']\n",
            "pred:  0 label:  0 ['people', 'will', 'prefer', 'to', 'instead', 'of', 'taking', '.']\n",
            "pred:  1 label:  1 ['as', 'of', '30', 'july', '2020', 'there', 'are', '477', 'active', 'cases', 'in', 'district', '.', 'for', 'details', 'kindly', 'contact', 'district', 'control', '.']\n",
            "pred:  0 label:  0 ['post', 'shared', 'more', 'than', 'times', 'on', 'facebook', 'during', 'novel', 'coronavirus', 'pandemic', 'says', 'bill', 'gates', 'wants', '?', '?', '?', 'digital', 'tattoos', '?', '?', 'to', 'check', 'who', 'has', 'been', 'tested', 'and', 'asks', 'if', 'it', 'would', 'be', '?', '?', '?', 'like', 'holocaust', 'victims', 'have', '.', '?', '?']\n",
            "pred:  0 label:  0 ['says', 'dr', '.', 'anthony', 'fauci', '?', '?', '?', 'stands', 'to', 'lose', '100', 'million', 'dollars', 'on', 'bill', 'gates', '?', '?', '?', 'vaccine', '?', '?', '?', 'for', 'coronavirus', '.', 'because', 'he', 'invested', 'in', 'this', 'vaccine', '.', '?', '?']\n",
            "pred:  0 label:  0 ['a', 'video', 'removed', 'from', 'in', 'which', 'a', 'woman', 'named', 'pilar', 'baselga', 'assures', 'that', 'carlos', 'iii', 'health', 'institute', 'of', 'madrid', 'has', 'said', 'in', 'a', 'report', 'that', 'coronavirus', 'pandemic', 'has', 'ended', '.']\n",
            "pred:  1 label:  1 ['hospitalist', 'compensation', '3', 'of', 'every', '10', 'are', 'women', '.', '.', '.']\n",
            "pred:  1 label:  1 ['a', 'mobile', 'testing', 'unit', 'went', 'to', 'airport', 'site', 'and', 'all', 'staff', 'who', 'work', 'there', 'have', 'now', 'been', 'tested', '.', 'mobile', 'testing', 'went', 'to', 'homes', 'of', 'symptomatic', 'people', 'who', 'work', 'in', 'mount', 'wellington', 'site', 'yesterday', '.', 'further', 'testing', 'is', 'underway', 'who', 'work', 'on', 'different', 'shifts', '.']\n",
            "pred:  1 label:  1 ['late', 'night', 'update', '.', 'states', 'have', 'disclosed', 'that', 'people', 'have', 'been', 'tested', '.', 'a', 'major', 'as', 'private', 'laboratories', 'begin', 'to', 'conduct', 'a', 'greater', 'percentage', 'of', 'tests', 'lose', 'them', 'from', 'state', 'data', '.', 'hoping', 'that', 'companies', 'and', 'states', 'opt', 'for', 'greater', 'transparency', '.']\n",
            "pred:  0 label:  0 ['group', 'founder', 'jack', 'ma', 'gave', 'million', 'to', 'a', 'filipino', 'student', 'who', 'developed', 'a', 'vaccine', 'for', 'novel', 'coronavirus', '.', '_']\n",
            "pred:  0 label:  0 ['goats', 'in', 'ajmer', 'rajasthan', 'were', 'found', 'to', 'be', 'positive', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'spelled', 'backwards', '=', 'a', 'coincidence', '?', '.']\n",
            "pred:  0 label:  0 ['pandemic', 'reveals', 'mayor', 'of', 'to', 'be', 'most', 'competent', 'us', 'politician', '.']\n",
            "pred:  1 label:  1 ['cdc', 'and', 'remind', 'you', 'that', 'wearing', 'cloth', 'face', 'coverings', 'in', 'public', 'settings', 'can', 'help', 'of', '.', 'learn', 'more', 'about', 'cloth', 'face', '.', '.', '.']\n",
            "pred:  0 label:  0 ['if', 'you', 'ask', 'amazon', '?', '?', '?', 's', 'alexa', 'about', 'origin', 'of', 'coronavirus', 'pandemic', ',', 'it', 'says', '?', '?', '?', 'government', 'planned', '?', '?', 'it', '.']\n",
            "pred:  1 label:  1 ['recovery', 'rate', 'improves', 'to', '77', '.', 'as', 'on', 'september', '13', '2020', 'improvement', 'in', 'recovery', 'rate', 'since', 'initiation', 'on', 'march', '25', '2020', 'via', '.']\n",
            "pred:  0 label:  0 ['vaccine', 'trial', 'killed', '7', 'children', 'in', 'senegal', '.']\n",
            "pred:  0 label:  0 ['french', 'state', 'gives', 'more', 'benefit', 'to', 'people', 'who', 'were', 'unemployed', 'before', 'than', 'it', 'does', 'for', 'people', 'who', 'are', 'partially', 'unemployed', 'due', 'to', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'states', 'reported', 'tests', 'cases', 'and', '449', 'deaths', '.', '.']\n",
            "pred:  1 label:  1 ['insurers', 'and', 'states', 'look', 'to', 'reduce', 'premiums', 'during', '.', '.']\n",
            "pred:  0 label:  0 ['politifact', 'will', 'keep', 'fact-checking', 'hoaxes', 'about', 'as', 'they', 'surface', '—', 'but', 'we', 'want', 'our', 'readers', 'to', 'be', 'prepared', 'too', '.', 'here', 'are', 'our', '7', 'tips', 'to', 'avoid', 'pandemic', 'misinformation', '.', '.', '.']\n",
            "pred:  1 label:  1 ['there', 'are', '40', 'people', 'isolating', 'in', 'auckland', 'quarantine', 'facility', 'from', 'community', 'which', 'includes', '17', 'people', 'who', 'have', 'tested', 'positive', 'for', 'and', 'their', 'household', 'contacts', '.']\n",
            "pred:  1 label:  1 ['shout', 'out', 'to', 'colorado', 'where', 'health', 'department', 'has', 'clearly', 'placed', 'emphasis', 'on', 'reporting', 'a', 'full', 'granular', 'dataset', 'on', 'outbreaks', '.', 'long-term', 'care', 'reporting', 'should', 'be', 'a', 'model', 'for', 'other', 'states', '.']\n",
            "pred:  1 label:  1 ['in', 'our', 'interconnected', 'world', 'if', 'people', 'in', 'and', 'middle-income', 'countries', 'miss', 'out', 'on', 'vaccines', 'virus', 'will', 'continue', 'to', 'kill', 'and', 'economic', 'recovery', 'will', 'be', 'delayed', '.', 'vaccine', 'nationalism', 'will', 'prolong', 'pandemic', 'not', 'shorten', '.']\n",
            "pred:  1 label:  1 ['number', 'of', 'people', 'states', 'report', 'to', 'be', 'hospitalized', 'with', 'continues', 'to', 'drop', '.', 'of', 'hospitalizations', 'are', 'currently', 'in', 'south', 'while', 'northeast', 'has', 'fallen', 'to', '.', '.']\n",
            "pred:  1 label:  1 ['we', 'will', 'continue', 'testing', 'in', 'community', 'as', 'part', 'of', 'our', 'ongoing', 'strategy', 'to', 'continue', 'with', 'elimination', 'of', '.', 'anyone', 'with', 'respiratory', 'symptoms', 'should', 'call', 'their', 'gp', 'or', 'healthline', 'on', '0800', '611', '116', 'to', 'get', 'advice', 'on', 'getting', 'a', 'test', '.', 'testing', 'is', 'free', 'in', 'nz', '.']\n",
            "pred:  1 label:  1 ['just', 'a', 'reminder', 'that', 'our', 'dms', 'are', 'open', 'for', 'testing', 'tidbits', 'you', 'may', 'have', 'for', 'your', 'state', '.', 'or', 'maybe', 'you', 'work', 'for', 'a', 'heath', 'department', 'or', 'big', 'testing', 'company', 'and', 'would', 'like', 'to', 'explain', 'going', 'on', '.', 'here', '.']\n",
            "pred:  0 label:  0 ['silver', 'solution', 'used', 'on', 'strains', 'of', 'coronavirus', 'eliminate', 'it', '.', 'kills', 'it', '.', 'deactivates', 'it', '.', '?', '?']\n",
            "pred:  0 label:  0 ['says', 'coronavirus', 'has', 'a', '?', '?', '?', '99', '.', 'recovery', 'rate', '?', '?', 'in', 'texas', '.']\n",
            "pred:  0 label:  0 ['maharashtra', 'governor', 'bhagat', 'singh', 'facilitated', 'travel', 'of', 'a', 'model', 'from', 'maharashtra', 'to', 'dehradun', 'during', 'lockdown', '.']\n",
            "pred:  0 label:  0 ['amid', 'coronavirus', 'crisis', 'streets', 'are', 'now', 'filled', 'with', 'discarded', 'printed', 'cash', 'as', 'seen', 'in', 'photos', 'circulating', 'on', 'social', 'media', '.']\n",
            "pred:  1 label:  1 ['13', 'states', 'are', 'reporting', 'hospitalized', 'patients', ';', '25', 'states', 'are', 'reporting', 'cumulative', 'hospitalizations', '.', 'these', 'numbers', 'are', 'not', 'alike', 'so', 'this', 'poses', 'a', 'considerable', 'challenge', 'to', 'tracking', 'this', 'important', 'data', '.']\n",
            "pred:  0 label:  0 ['holding', 'your', 'breath', 'for', '10', 'seconds', 'is', 'a', 'way', 'to', 'test', 'for', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['there', 'are', '16', 'previously', 'reported', 'cases', 'who', 'are', 'considered', 'to', 'have', 'recovered', 'all', 'community', 'cases', '.']\n",
            "pred:  1 label:  1 ['rt', 'has', 'potential', 'to', 'disrupt', 'health', 'services', '-', 'systems', 'becoming', 'overwhelmed', 'intervention', 'to', 'slow']\n",
            "pred:  0 label:  0 ['says', 'democrats', 'are', 'on', 'vacation', 'until', 'may', '4', 'and', '?', '?', '?', 'refuse', 'to', 'come', 'back', '?', '?', 'to', 'sign', 'a', 'bill', 'to', 'help', 'small', 'businesses', '.']\n",
            "pred:  0 label:  0 ['progressive', 'sugar', 'is', 'bad', 'for', 'you', ';', 'government', 'surveillance', 'is', 'necessary', 'to', 'keep', '.']\n",
            "pred:  0 label:  0 ['this', 'is', 'a', 'map', 'of', 'india', 'redesigned', 'by', 'an', 'american', 'ceo', 'where', 'he', 'marked', 'indian', 'states', 'population', 'which', 'is', 'almost', 'equal', 'to', 'population', 'of', 'some', 'of', 'countries', '.', 'he', 'tried', 'to', 'explain', 'to', 'his', 'employees', 'that', 'india', 'is', 'indirectly', 'handling', 'situation', 'of', 'so', 'many', 'countries', '.']\n",
            "pred:  0 label:  0 ['has', 'already', 'developed', 'a', 'vaccine', 'against']\n",
            "pred:  0 label:  0 ['obese', 'man', 'most', 'likely', 'to', 'die', 'first', 'according', 'to', 'family', 'sweepstake', '.']\n",
            "pred:  0 label:  0 ['a', 'photo', 'shows', 'rob', 'at', 'a', 'recent', 'michigan', 'protest', '.']\n",
            "pred:  1 label:  1 ['some', 'coronavirus', 'patients', 'are', 'experiencing', 'chronic', 'fatigue', '.', '.', '.']\n",
            "pred:  1 label:  1 ['india', 'tracker', 'on', '16', 'september', '2020', '08:00', '.', '.', '.', 'via', '.']\n",
            "pred:  1 label:  1 ['controls', 'at', 'borders', 'remain', 'for', 'those', 'entering', 'new', 'zealand', 'including', 'health', 'screening', 'and', 'testing', 'for', 'all', 'arrivals', 'and', 'mandatory', '14', 'day', 'managed', 'quarantine', 'or', 'isolation']\n",
            "pred:  0 label:  0 ['government', 'has', 'finally', 'and', 'will', 'be', 'giving', 'out', ',', '000', 'in', 'to', 'every', 'citizen']\n",
            "pred:  1 label:  1 ['6', 'cases', 'are', 'now', 'considered', 'recovered', 'bringing', 'number', 'of', 'active', 'cases', 'in', 'new', 'zealand', 'to', '18', '.', 'all', 'are', 'in', 'managed', 'isolation', 'or', 'quarantine', 'facilities', '.']\n",
            "pred:  0 label:  0 ['we', 'have', 'about', '8', 'to', '10', 'bodies', 'a', 'day', ',', 'but', 'one', 'time', 'we', 'had', '18', 'bodies', ',', '\"', 'says', 'head', 'of', 'one', 'of', 'few', 'crematoriums', 'on', 'bali', 'designated', 'to', 'deal', 'with', 'bodies', 'of', 'people', 'who', 'died', 'with', 'confirmed', 'or', 'suspected', 'cases', 'of', '.', 'report', 'by', '.', '.']\n",
            "pred:  0 label:  0 ['claim', 'that', 'saint', 'corona', 'has', 'always', 'been', 'a', 'patron', 'saint', 'against', 'epidemics', '.']\n",
            "pred:  1 label:  1 ['update', 'from', 'today', 'we', 'have', '1', 'new', 'case', 'of', 'to', 'report', 'in', 'managed', 'isolation', 'facilities', 'in', 'nz', 'it', 'has', 'been', '81', 'days', 'since', 'last', 'case', 'of', 'was', 'acquired', 'locally', 'from', 'an', 'unknown', 'source', '.', 'our', 'total', 'number', 'of', 'active', 'cases', 'is', '27', '.']\n",
            "pred:  1 label:  1 ['leads', 'rapid', 'response', 'task', 'force', ';', 'a', 'critical', 'black', 'digital', 'humanities', 'collective', 'that', 'connects', 'black', 'diasporic', 'communities', 'impacted', 'by', 'pandemic', 'to', 'their', 'data', '.', '”', 'they', 'have', 'a', 'bunch', 'of', 'worthwhile', 'projects', 'in', 'works', '.', '.']\n",
            "pred:  1 label:  1 ['adding', 'new', 'incidence', 'metric', 'has', 'changed', 'scores', 'of', '350', 'counties', 'and', 'our', 'coverage', 'to', '1800', 'counties', '.', '**', 'previously', 'many', 'counties', 'did', 'not', 'have', 'enough', 'data', 'for', 'us', 'to', 'calculate', 'a', 'risk', 'score', '.', 'it', 'is', 'critical', 'for', 'local', 'decision-makers', 'to', 'have', 'local-level', 'data', '.']\n",
            "pred:  0 label:  0 ['is', 'common', 'flu', 'and', 'cure', 'for', 'coronavirus', 'pneumonia', 'is', 'easy', 'and', 'cheap', '.']\n",
            "pred:  1 label:  1 ['equity', 'accuracy', 'and', 'quality', 'questions', 'surround', 'vaccine', '.', '.']\n",
            "pred:  0 label:  0 ['says', 'imposters', 'in', 'hazmat', 'suits', 'are', 'going', 'door', 'to', 'door', 'in', 'stockton', 'california', 'saying', 'they', 'are', 'checking', 'residents', 'for', 'fever', 'or', 'but', 'they', 'enter', 'your', 'home', 'and', 'physically', 'attempt', 'robbery', '.', '”']\n",
            "pred:  1 label:  1 ['of', 'cases', 'in', 'have', 'unknown', 'sources', 'of', 'infection', 'this', 'is', 'normal', 'for', 'a', 'respiratory', 'virus', ';', 'suggests', 'ongoing', 'community', 'transmission', 'in', '~', 'ncdc', 'director', 'general', 'at', '.']\n",
            "pred:  0 label:  0 ['a', 'viral', 'text', 'based', 'image', 'post', 'claiming', 'that', '5', ',', '000', 'from', 'italy', 'were', 'infected', 'by', 'mobile', 'phone', 'covers', ',', 'quoting', 'findings', 'were', 'from', 'a', 'research', 'without', 'specifying', 'any', 'details', 'or', 'name', 'of', 'research', '.', 'post', 'advises', 'to', 'abandon', 'phone', 'covers', '.']\n",
            "pred:  0 label:  0 ['we', 'cannot', 'return', 'to', 'ed', 'sheeran', ',', 'warn', 'experts', '.', '.']\n",
            "pred:  1 label:  1 ['health', 'ministry', 'issues', 'updated', 'advisory', 'on', 'testing', '.', 'simplified', 'testing', 'procedure', ';', 'testing', 'for', 'first', 'time', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'only', 'of', 'people', 'actually', 'died', 'from', '.', '?', '?', 'others', '?', '?', '?', 'died', 'from', 'other', 'reasons', '.', '?', '?']\n",
            "pred:  0 label:  0 ['olive', 'garden', 'employees', 'are', 'allowed', 'to', 'wear', 'black', 'lives', 'matter', 'masks', 'not', 'american', 'flag', '.']\n",
            "pred:  1 label:  1 ['update', 'from', 'today', 'there', 'is', '1', 'new', 'case', 'of', 'to', 'report', 'in', 'managed', 'isolation', 'in', 'new', 'zealand', '.', 'there', 'continue', 'to', 'be', 'no', 'cases', 'of', 'in', 'community', '.', 'it', 'has', 'been', '66', 'days', 'since', 'last', 'case', 'of', 'that', 'was', 'acquired', 'locally', 'from', 'an', 'unknown', 'source', '.']\n",
            "pred:  0 label:  0 ['g-7', 'countries', 'have', 'asked', 'modi', 'to', 'lead', 'them', 'in', 'preventing', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['minister', 'of', 'health', 'claims', 'that', 'only', 'one', 'case', 'of', 'registered', 'in', 'lithuania', '.']\n",
            "pred:  0 label:  0 ['july', '2020', 'has', 'less', 'deaths', 'than', 'july', '2019', 'despite', 'pandemic']\n",
            "pred:  1 label:  1 ['rt', 'during', 'pandemic', 'it', 'is', 'even', 'more', 'important', 'to', 'find', 'innovative', 'solutions', 'to', 'ensure', 'that', 'access', 'to']\n",
            "pred:  0 label:  0 ['this', 'is', 'rahul', 'gandhi', ';', 'priyanka', 'gandhi', 'breaking', 'lock-down', 'regulations', 'saying', 'they', 'need', 'to', 'visit', 'friends', '.']\n",
            "pred:  1 label:  1 ['5', 'andhra', 'pradesh', 'karnataka', 'uttar', 'pradesh', ';', 'tamil', 'nadu', 'account', 'for', 'of', 'total', 'active', 'cases', 'in', 'secretary', '.']\n",
            "pred:  1 label:  1 ['medical', 'entomologist', 'janet', 'mcallister', 'and', 'other', 'cdc', 'responders', 'had', 'to', 'adapt', 'to', 'changing', 'needs', 'and', 'conditions', 'when', 'responding', 'to', 'outbreaks', 'in', 'wisconsin', '.', 'learn', 'about', 'her', '.', '.', '.']\n",
            "pred:  1 label:  1 ['on', 'a', 'bigger', 'than', 'average', 'testing', 'day', 'newly', 'reported', 'cases', 'were', 'over', '25k', 'today', '.', '.']\n",
            "pred:  1 label:  1 ['on', 'a', 'steep', 'rise', 'india', 'conducts', 'more', 'than', '4', '.', '23', 'cr', 'tests', 'of', 'total', 'cases', 'recorded', 'in', 'just', '3', 'ap', 'karnataka', '.']\n",
            "pred:  0 label:  0 ['peru', 'has', 'just', 'nationalized', 'hospitals', 'and', 'clinics', '.']\n",
            "pred:  0 label:  0 ['a', 'nurse', 'claims', 'that', 'delhi', 'govt', 'has', 'stopped', 'funding', 'for', 'hotel', 'stay', 'for', 'doctors', 'and', 'nurses', 'treating', '.']\n",
            "pred:  0 label:  0 ['terminally', 'ill', 'patients', 'resentful', 'not', 'dying', 'from', '.']\n",
            "pred:  0 label:  0 ['lot', 'of', 'news', 'coming', 'in', 'of', 'irregularities', 'in', 'treatment', 'of', 'patients', 'in', 'several', 'major', 'hospitals', 'across', '.', 'if', 'you', 'have', 'experienced', 'something', 'unpleasant', 'or', 'are', 'in', 'a', 'situation', 'that', 'calls', 'for', 'immediate', 'attention', ',', 'please', 'feel', 'free', 'to', 'reach', 'out', '.']\n",
            "pred:  0 label:  0 ['physicians', 'in', 'video', 'seen', 'by', 'millions', 'are', 'associated', 'with', 'a', 'group', 'called', 'frontline', 'doctors', 'which', 'advocates', 'against', 'official', 'narratives', 'of', '.', 'all', 'of', 'physicians', 'have', 'a', 'history', 'of', 'making', 'unproven', 'or', 'conspiratorial', 'medical', 'claims', '.', '.']\n",
            "pred:  1 label:  1 ['vaccine', 'availability', 'has', 'increased', 'significantly', 'in', 'most', 'cold', 'chain', 'points', 'since', 'introduction', 'of', 'along', 'with', 'a', 'significant', 'reduction', 'in', 'vaccine', 'wastage', '.', '.']\n",
            "pred:  0 label:  0 ['breaking', 'who', 'says', 'vaccine', 'for', 'is', 'ready', '.', 'just', 'go', 'and', 'buy', 'a', 'bmw', 'z', 'class']\n",
            "pred:  1 label:  1 ['everyone', 'who', 'was', 'in', 'bus', 'from', 'auckland', 'airport', 'to', 'ibis', 'hotel', 'is', 'being', 'followed', 'up', 'and', 'are', 'already', 'in', 'managed', 'isolation', '.', 'driver', 'who', 'was', 'wearing', 'ppe', 'has', 'also', 'been', 'isolated', 'as', 'a', 'potential', 'close', 'contact', 'and', 'will', 'be', 'tested', '.']\n",
            "pred:  0 label:  0 ['not', 'a', 'single', 'muslim', 'died', 'of', 'coronavirus', 'in', 'china', '.']\n",
            "pred:  1 label:  1 ['added', 'an', 'important', 'new', 'column', 'of', 'hospitalizations', '.', 'only', '11', 'states', 'currently', 'report', 'it', 'but', 'we', 'started', 'tracking', 'it', 'today', '.', 'state', 'data', 'pages', 'see', 'we', 'have', 'some', 'design', 'changes', 'to', 'make', 'but', 'we', 'wanted', 'to', 'get', 'this', 'urgent', 'information', 'out', 'there', 'right', 'away', '.', ')']\n",
            "pred:  1 label:  1 ['note', 'that', 'oklahoma', 'reported', 'positive', 'tests', 'but', 'not', 'negative', 'or', 'total', 'tests', 'today', '.', 'this', 'almost', 'certainly', 'does', 'not', 'mean', 'ok', 'has', 'a', 'positive', 'rate', '.', 'a', 'reporting', 'delay', '.']\n",
            "pred:  0 label:  0 ['says', 'gov', 'tony', 'evers', '?', '?', '?', 'administration', 'deemed', 'dentists', 'non-essential']\n",
            "pred:  0 label:  0 ['nashville', 'man', 'devastated', 'when', 'girlfriend', 'tells', 'him', 'she', 'wants', 'to', 'start', 'social', 'distancing', '.']\n",
            "pred:  0 label:  0 ['health', 'insurance', 'industry', 'has', '?', '?', '?', 'agreed', 'to', 'waive', 'all', 'co-payments', 'for', 'coronavirus', 'treatments', '.', '?', '?']\n",
            "pred:  1 label:  1 ['younger', 'patients', 'are', 'harder', 'to', 'track', 'than', 'their', 'elders', '.', '.']\n",
            "pred:  1 label:  1 ['california', 'missouri', 'oklahoma', 'and', 'north', 'dakota', 'all', 'set', 'new', 'records', 'for', 'cases', 'today', '.', '.']\n",
            "pred:  1 label:  1 ['colleges', 'drop', 'courses', 'including', 'languages', 'and', 'maths', 'amid', 'financial', 'pressure', '.']\n",
            "pred:  0 label:  0 ['cdc', 'recommends', 'men', 'shave', 'their', 'beards', 'to', 'protect', 'against', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['multiple', 'allegations', 'on', 'bill', 'drafted', 'to', 'turkish', 'congress', 'named', 'h', '.', 'r', '.', '.']\n",
            "pred:  0 label:  0 ['obama', 'administration', 'made', 'a', 'decision', 'on', 'testing', 'that', 'turned', 'out', 'to', 'be', 'very', 'detrimental', 'to', 'what', 'on', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['antonio', 'guterres', 'said', 'pandemic', 'has', 'shown', 'how', 'we', 'and', 'that', 'in', 'some', 'aspects', 'are', 'moving', '25', 'years', '.']\n",
            "pred:  0 label:  0 ['fringe', 'japanese', 'religion', 'that', 'claims', 'it', 'can', 'cure', '.', 'new', 'york', 'times', '.']\n",
            "pred:  1 label:  1 ['announced', 'they', 'are', 'launching', 'a', 'of', 'large', 'clinical', 'trials', 'to', 'test', 'new', 'approaches', 'to', 'treating', '.', 'effort', 'aims', 'to', 'develop', 'a', 'public-private', 'coordinated', 'approach', 'to', 'prioritizing', 'and', 'speeding', 'up', 'development', 'of', 'treatments', 'and', 'vaccines', '.', 'via']\n",
            "pred:  0 label:  0 ['government', 'insists', 'there', 'will', 'be', 'no', 'u-turn', 'on', 'schools', 'reopening', ',', 'until', 'after', 'everyone', 'has', 'spent', 'on', 'new', 'uniforms', '.', '.']\n",
            "pred:  0 label:  0 ['unicef', 'releases', 'coronavirus', 'prevention', 'guidelines', '.']\n",
            "pred:  1 label:  1 ['but', 'texas', 'north', 'carolina', 'and', 'south', 'carolina', 'also', 'saw', 'spikes', 'in', 'hospitalizations', '.', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'states', 'reported', 'tests', 'and', 'cases', '.', 'death', 'toll', 'today', 'is', '1147', '.', 'after', 'lower', 'reported', 'deaths', 'over', 'weekend', 'today', 'we', 'see', 'totals', 'back', 'over', '1000', '.', '.']\n",
            "pred:  0 label:  0 ['cruise', 'lines', 'now', 'offer', 'deluxe', 'coronavirus', 'upgrade', 'over', 'basic', 'norovirus', 'package', '.']\n",
            "pred:  1 label:  1 ['another', 'ca', 'note', '.', 'state', 'had', 'been', 'reporting', 'irregularly', 'so', 'we', 'were', 'relying', 'on', 'high-quality', 'media', 'outlets', 'like', 'who', 'have', 'an', 'excellent', 'tracker', '.', '.', 'now', 'that', 'state', 'is', 'reporting', 'more', 'regularly', 'we', 'will', 'be', 'relying', 'again', 'on', 'state', 'data', '.']\n",
            "pred:  1 label:  1 ['recommendations', 'for', 'treating', 'colorectal', 'cancer', 'during', 'pandemic', '.', '.']\n",
            "pred:  0 label:  0 ['prolonged', 'use', 'of', 'face', 'masks', 'causes', 'o2', 'deficiency', 'or', 'co2', 'intoxication', '.']\n",
            "pred:  1 label:  1 ['how', 'are', 'you', 'keeping', 'current', 'on', 'guidelines', 'and', 'updates', 'during', '?', 'tell', 'us', 'about', 'it', 'in', 'this', 'confidential', 'survey', '.', 'no', 'name', 'or', 'email', 'required', '.', '.', '.']\n",
            "pred:  1 label:  1 ['a', 'reminder', 'that', 'auckland', 'will', 'be', 'moving', 'to', 'alert', 'level', '2', '.', '5', 'at', 'tonight', '.', 'social', 'gatherings', 'are', 'limited', 'to', '10', '-', 'this', 'includes', 'everything', 'from', 'birthdays', 'family', 'gatherings', 'even', 'friends', 'and', 'neighbours', '.']\n",
            "pred:  1 label:  1 ['this', 'week', '19', 'states', 'posted', 'their', 'highest', 'number', 'of', 'new', 'cases', 'reported', 'in', 'a', 'single', 'day', '.', 'three', 'of', 'them', 'hit', 'those', 'record', 'highs', 'today', '.', '.']\n",
            "pred:  0 label:  0 ['government', 'must', 'have', 'planned', 'coronavirus', 'pandemic', 'because', 'coronavirus', 'relief', 'bill', 'had', 'been', 'introduced', 'three', 'years', 'earlier', '.']\n",
            "pred:  0 label:  0 ['joe', 'biden', 'has', 'on', 'several', 'occasions', 'said', 'he', 'was', 'first', 'person', 'to', 'call', 'for', 'invoking', 'defense', 'production', 'act', 'in', 'response', 'to', 'coronavirus', '.', 'we', 'looked', 'at', 'timeline', '.', 'he', '.', '.']\n",
            "pred:  0 label:  0 ['getting', 'a', 'flu', 'shot', 'increases', 'risk', 'of', 'coronavirus', 'by', '.']\n",
            "pred:  1 label:  1 ['chronic', 'has', 'long-lasting', 'health', 'effects', '.', '.']\n",
            "pred:  0 label:  0 ['scientists', 'at', 'research', 'institute', 'have', 'developed', 'a', 'vaccine', 'for', '.', '_']\n",
            "pred:  1 label:  1 ['rt', '.', 'launches', 'new', 'lab', 'for', 'testing', 'to', 'help', 'passengers', 'coming', 'to', 'delhi', 'from', 'international']\n",
            "pred:  1 label:  1 ['rt', 'update', '!', 'deaths', 'per', 'million', 'population', 'in', 'india', 'is', 'one', 'of', 'lowest', 'in', 'world', '.', 'global']\n",
            "pred:  0 label:  0 ['a', 'photo', 'of', 'a', 'queue', 'of', 'buses', 'in', 'india', 'has', 'been', 'shared', 'thousands', 'of', 'times', 'on', 'facebook', 'and', 'twitter', 'alongside', 'a', 'claim', 'they', 'were', 'organised', 'by', 'a', 'leading', 'opposition', 'politician', 'to', 'transport', 'migrant', 'workers', 'who', 'were', 'left', 'stranded', 'after', 'a', 'nationwide', 'coronavirus', 'lockdown', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'suspects', 'to', 'be', 'quarantined', 'on', 'love', 'island', '.', '.']\n",
            "pred:  0 label:  0 ['emergency', 'advice', 'you', 'need', 'to', 'hear', '.']\n",
            "pred:  0 label:  0 ['president', 'donald', 'trump', 'announced', 'commercialization', 'of', 'a', 'vaccine', '.']\n",
            "pred:  1 label:  1 ['am', 'pleased', 'to', 'report', 'that', 'state', 'is', 'now', 'aligned', 'with', 'federal', 'government', 'on', 'response', 'has', 'started', 'testing', 'as', 'well', 'as', 'reporting', 'on', 'national', 'dash', '~', 'hmh', 'at', '.', '.']\n",
            "pred:  1 label:  1 ['india', 'overtakes', 'usa', 'to', 'become', 'no', '.', '1', 'in', 'terms', 'of', 'global', 'recoveries', '.', 'india', 'accounts', 'for', 'of', 'global', 'recoveries', '.', '.', '.']\n",
            "pred:  0 label:  0 ['nigerians', 'are', '?', '?', '?', 'burning', 'chinese', 'stores', '.', '?', '?']\n",
            "pred:  1 label:  1 ['131', 'new', 'cases', 'of', ';', 'akwa', 'cross', 'confirmed', 'discharged', '1093', 'deaths', '.']\n",
            "pred:  1 label:  1 ['rt', 'for', 'people', 'in', 'recovery', 'can', 'make', 'it', 'difficult', 'to', 'get', 'support', '.', 'find', 'information', 'on', 'and', 'substance', 'use']\n",
            "pred:  0 label:  0 ['after', 'amit', 'shah', 'tested', 'positive', 'some', 'people', 'took', 'out', 'his', 'funeral', '.']\n",
            "pred:  0 label:  0 ['photos', 'show', 'part', 'of', 'hospital', 'in', 'wuhan', 'built', 'in', 'five', 'days', '.']\n",
            "pred:  1 label:  1 ['repeated', 'laboratory', 'testing', 'combined', 'with', 'infection', 'prevention', 'control', 'interventions', 'likely', 'contributed', 'to', 'decreases', 'in', 'spread', 'in', '26', 'detroit', 'skilled', 'nursing', 'facilities', '.', 'read', 'more', 'in', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'set', 'up', 'programs', 'in', '47', 'vulnerable', 'countries', 'as', 'a', 'way', 'to', 'protect', 'against', 'something', 'like', 'coronavirus', ',', 'exactly', '.', 'do', 'you', 'know', 'that', 'trump', 'closed', '37', 'of', 'them', '?', '?', '?']\n",
            "pred:  1 label:  1 ['pandemic', 'has', 'taken', 'a', 'huge', 'toll', 'on', 'lives', ';', 'livelihoods', 'disrupting', 'health', 'systems', 'economies', ';', 'societies', '.', 'even', 'countries', 'with', 'advanced', 'health', 'systems', ';', 'powerful', 'economies', 'have', 'been']\n",
            "pred:  1 label:  1 ['there', 'were', '332', 'reported', 'deaths', 'today', '.', '7-day', 'average', 'in', 'deaths', 'remains', 'about', 'unchanged', '.', 'why', 'are', 'deaths', 'falling', 'or', 'holding', 'steady', 'while', 'cases', 'rise', '?', 'probably', 'for', 'a', 'few', '.', '.']\n",
            "pred:  1 label:  1 ['for', 'example', 'if', 'got', 'a', 'cold', 'cough', 'or', 'other', 'respiratory', 'symptoms', 'and', 'virus', 'is', 'picked', 'up', 'as', 'a', 'weak', 'positive', 'in', 'a', 'pcr', 'test', '.', 'in', 'all', 'these', 'cases', 'test', 'result', 'is', 'often', 'a', 'and', 'ct', 'value', 'of', 'pcr', 'test', 'suggests', 'an', 'old', 'infection', '.']\n",
            "pred:  1 label:  1 ['attend', 'coca', 'call', 'on', 'telehealth', ';', 'health', 'considerations', 'for', 'addressing', 'health', 'disparities', 'during', 'pandemic', 'at', '2pm', 'et', '.', 'learn', '.', '.', '.']\n",
            "pred:  1 label:  1 ['from', 'friday', 'pubs', 'bars', 'and', 'restaurants', 'will', 'have', 'to', 'close', 'at', '10pm', '.', 'first', 'minister', 'of', 'scotland', 'nicola', 'sturgeon', 'announces', 'a', 'curfew', 'for', 'pubs', '.', 'updates', 'on', 'latest', '.', '.']\n",
            "pred:  1 label:  1 ['medical', 'groups', 'say', 'in', 'unison', 'wear', 'a', 'mask', '.', '.', 'via']\n",
            "pred:  0 label:  0 ['whatsapp', 'text', 'that', 'says', 'chinese', 'created', 'coronavirus', 'and', 'that', 'main', 'cities', 'in', 'china', 'did', 'not', 'suffer', 'from', 'pandemic', '.']\n",
            "pred:  0 label:  0 ['house', 'resolution', 'provides', 'funds', 'to', 'forcibly', 'removed', 'people', 'with', 'from', 'their', 'homes', 'as', 'dr', '.', 'rashid', 'buttar', 'said', '.']\n",
            "pred:  1 label:  1 ['total', 'deaths', 'remain', 'highest', 'in', 'northeast', 'with', 'nearly', 'a', 'substantial', 'undercount', 'as', 'new', 'york', 'reports', 'only', 'deaths', 'within', 'long-term', 'care', 'facilities', 'not', 'those', 'linked', 'to', 'such', 'facilities', 'but', 'occurring', 'in', 'hospitals', '.', '.']\n",
            "pred:  0 label:  0 ['.', 'please', 'ask', 'your', 'officers', 'to', 'wear', 'masks', 'properly', 'while', 'they', 'fine', 'others', 'for', '.']\n",
            "pred:  0 label:  0 ['gov', '.', 'andrew', 'cuomo', '?', '?', '?', 'was', 'simply', 'saying', 'if', 'we', 'can', 'share', '20', 'percent', 'of', 'your', 'excess', ',', 'your', 'ventilators', ',', 'to', 'help', 'people', 'in', 'other', 'parts', 'of', 'state', 'on', 'a', 'voluntary', 'basis', ',', 'that', 'would', 'be', 'great', '.', 'of', 'course', ',', 'there', 'was', 'a', 'reaction', 'to', 'that', ',', 'which', 'was', 'not', 'positive', '.', '\"']\n",
            "pred:  0 label:  0 ['a', 'video', 'of', 'dozens', 'of', 'people', 'praying', 'outdoors', 'shows', 'italians', 'praying', 'during', 'novel', 'coronavirus', 'pandemic', '.']\n",
            "pred:  1 label:  1 ['409', 'new', 'cases', 'of', ';', '85', '22', '16', '14', '14', '14', '5', '5', '4', '4', '2', '2', '2', 'total', 'confirmed', 'cases', 'discharged', '382', 'deaths', '.']\n",
            "pred:  0 label:  0 ['for', 'its', 'must-read', 'coronavirus', 'coverage', 'atlantic', 'is', 'rewarded', 'with', 'a', 'huge', 'surge', 'of', 'digital', 'subscriptions', '.', '.']\n",
            "pred:  1 label:  1 ['yesterday', 'our', 'laboratories', 'completed', '2403', 'tests', '.', 'total', 'number', 'of', 'tests', 'completed', 'in', 'new', 'zealand', 'to', 'date', 'is', '.']\n",
            "pred:  1 label:  1 ['as', 'noted', 'earlier', 'we', 'had', 'to', 'do', 'some', 'work', 'with', 'numbers', 'as', 'comprehensive', 'set', 'had', 'not', 'come', 'in', 'by', 'time', 'we', 'published', 'our', 'daily', 'update', '.', 'make', 'sure', 'you', 'check', 'out', 'our', 'solution', 'is', 'still', 'not', '.']\n",
            "pred:  0 label:  0 ['durango', 'police', 'make', 'an', 'extremely', 'unusual', 'arrest', '.']\n",
            "pred:  1 label:  1 ['as', 'of', 'today', 'for', 'cases', 'in', 'india', 'is', '76', '.', '.', '.']\n",
            "pred:  1 label:  1 ['india', 'tracker', 'on', '30', 'august', '2020', '08:00', '.', '.', '.', 'via', '.']\n",
            "pred:  1 label:  1 ['yesterday', 'our', 'laboratories', 'completed', 'tests', 'bringing', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  1 label:  1 ['doctors', 'and', 'healthcare', 'workers', 'need', 'ppe', 'now', 'more', 'than', 'ever', '.', 'is', 'determined', 'to', 'get', 'ppe', 'supplies', 'to', 'healthcare', 'workers', 'we', 'love', 'what', 'whether', 'it', 'be', 'donating', 'funds', 'or', 'donating', 'your', 'own', 'supplies', 'see', 'how', 'you', 'can', 'help', 'at', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'update', 'records', 'more', 'than', 'recoveries', 'for', 'two', 'days', 'in', 'a', 'row', 'high', 'levels', 'of', 'recovery']\n",
            "pred:  0 label:  0 ['writer', 'sees', 'book', 'sales', 'soar', 'as', 'coronavirus', 'spreads', '.']\n",
            "pred:  1 label:  1 ['we', 'want', 'to', 'note', 'an', 'organizational', 'change', '.', 'has', 'returned', 'to', 'running', 'his', 'company', '.', 'we', 'are', 'so', 'grateful', 'to', 'him', 'for', 'an', 'astonishing', 'amount', 'of', 'work', 'getting', 'this', 'project', 'off', 'ground', '.', 'please', 'send', 'future', 'complaints', 'to', 'me', '.', 'our', 'team', '.']\n",
            "pred:  1 label:  1 ['rt', '.', 'is', 'leveraging', 'all', 'available', 'surveillance', 'systems', 'to', 'monitor', 'and', 'protect', 'at', 'risk', 'populations', '.', 'we', 'are']\n",
            "pred:  1 label:  1 ['there', 'remains', 'one', 'person', 'receiving', 'care', 'for', ';', 'they', 'are', 'in', 'middlemore', 'and', 'are', 'not', 'in', 'icu', '.', 'yesterday', 'our', 'laboratories', 'completed', 'tests', 'bringing', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  0 label:  0 ['a', 'viral', 'video', 'shows', 'a', 'group', 'of', 'doctors', 'airing', 'unproven', 'conspiracy', 'theories', 'about', 'coronavirus', '.', 'we', 'one', 'of', 'most', 'inaccurate', 'claims', '.', '.']\n",
            "pred:  0 label:  0 ['italy', 'is', 'burying', 'corona', 'victims', 'in', 'mass', 'graves', '.']\n",
            "pred:  0 label:  0 ['comments', 'section', 'quarantined', 'for', 'outbreak', 'of', 'racism', '.', '.']\n",
            "pred:  0 label:  0 ['world', 'health', 'organization', 'released', 'a', 'list', 'of', 'seven', 'habits', 'as', 'biggest', 'habits', '.']\n",
            "pred:  0 label:  0 ['french', 'doctors', 'protest', 'after', 'discovering', 'pandemic', 'is', 'a', 'fraud', '.']\n",
            "pred:  1 label:  1 ['rt', 'increase', 'in', 'testing', 'has', 'led', 'to', 'a', 'surge', 'in', 'total', 'number', 'of', 'recovered', 'patients', 'which', 'is', 'nearly', '30', 'lakh']\n",
            "pred:  1 label:  1 ['healthcare', 'learn', 'how', 'telehealth', 'technology', 'can', 'help', 'you', 'safely', 'provide', 'necessary', 'care', 'to', 'patients', 'during', 'pandemic', '.', '.', '.']\n",
            "pred:  0 label:  0 ['bill', 'gates', 'tried', 'to', 'get', 'a', 'patent', 'for', 'masks', '.']\n",
            "pred:  1 label:  1 ['on', 'a', 'steady', 'march', 'india', 'conducts', 'a', 'total', 'of', 'nearly', '3', '.', '6', 'crore', 'tests', '.', 'tests', 'per', 'million', 'scale', 'a', 'new', 'peak', 'of', '.', '.', '.']\n",
            "pred:  1 label:  1 ['of', '2159', 'people', 'who', 'left', 'managed', 'isolation', 'facilities', 'between', 'june', '9-16', '1313', 'people', 'have', 'been', 'contacted', 'and', 'have', 'tested', 'negative', 'for', ';', '800', 'of', 'those', 'were', 'tested', 'before', 'leaving', 'managed', 'isolation', 'and', 'remaining', '513', 'were', 'tested', 'after', 'departure', 'from', 'facility', '.']\n",
            "pred:  0 label:  0 ['mycoplasma', 'pneumonia', 'can', 'be', 'treated', 'with', 'quinolones', '.', 'first', 'quinolone', 'discovered', 'was', 'chloroquine', '.', 'mycoplasma', 'pneumonia', 'is', 'a', 'rare', 'but', 'important', 'cause', 'of', 'transverse', 'myelitis', '.', '.']\n",
            "pred:  1 label:  1 ['a', 'new', 'report', 'in', 'shows', 'that', 'rate', 'of', 'among', 'american', 'indian', 'and', 'alaska', 'native', 'people', 'was', '3', '.', '5', 'times', 'that', 'of', 'non-hispanic', 'white', 'persons', 'in', 'first', '6', 'months', 'of', 'pandemic', 'in', '23', 'states', '.', 'learn', '.', '.', '.']\n",
            "pred:  0 label:  0 ['italy', 'finds', 'home', 'remedy', 'for', '/', 'aspirin', 'dissolved', 'with', 'lemon', 'juice', 'boiled', 'with', 'honey', ',', 'it', 'is', 'taken', 'hot', '.']\n",
            "pred:  1 label:  1 ['second', 'case', 'is', 'a', 'woman', 'in', 'her', '30s', 'who', 'arrived', 'in', 'new', 'zealand', 'on', 'july', '28', 'from', 'los', 'angeles', '.', 'she', 'has', 'been', 'staying', 'at', 'in', 'auckland', 'and', 'tested', 'positive', 'for', 'as', 'part', 'of', 'routine', 'testing', 'around', 'day', 'three', 'of', 'her', 'stay', 'in', 'managed', 'isolation', '.']\n",
            "pred:  1 label:  1 ['for', 'more', 'information', 'on', 'symptoms', 'of', 'and', 'what', 'to', 'watch', 'for', '.', '.']\n",
            "pred:  0 label:  0 ['people', 'being', 'sent', 'so', 'far', 'for', 'tests', ',', 'they', 'have', 'to', 'quarantine', 'for', '14', 'days', 'upon', 'return', '.', '.']\n",
            "pred:  0 label:  0 ['hindus', 'cast', 'their', 'idols', 'in', 'sea', 'because', 'they', 'did', 'not', 'defend', 'them', 'against', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['says', 'state', 'secretary', 'says', 'go', 'to', 'hospitals', 'or', 'clinics', 'now', 'tourism', 'secretary', 'says', 'recreate', '.', '”']\n",
            "pred:  0 label:  0 ['fenugreek', 'seed', 'tape', 'can', 'cure']\n",
            "pred:  0 label:  0 ['a', 'muslim', 'woman', 'is', 'shouting', 'at', 'doctors', 'who', 'have', 'come', 'for', 'coronavirus', 'inspection', 'at', 'one', 'of', 'places', 'in', 'india', '.']\n",
            "pred:  1 label:  1 ['care', 'home', 'staff', 'in', 'hotspots', 'are', 'waiting', 'over', 'a', 'week', 'for', 'test', 'results', 'prompting', 'concerns', 'that', 'system', 'cannot', 'cope', 'with', 'increased', 'demand', '.']\n",
            "pred:  1 label:  1 ['as', 'at', '8:35', 'pm', '26th', 'march', 'there', 'are', '65', 'confirmed', 'cases', '3', 'discharged', '1', 'death', 'for', 'a', 'breakdown', 'of', 'cases', 'by', 'states', 'in', 'real', 'time', 'please', 'see', '.', 'currently', ';', '44', '11', '3', '1', '1', '1', '2']\n",
            "pred:  0 label:  0 ['novel', 'coronavirus', 'was', 'made', 'in', 'a', 'lab', 'and', 'is', 'now', 'spread', 'as', 'a', 'way', 'to', 'force', 'vaccinate', 'people', '.', '_']\n",
            "pred:  0 label:  0 ['a', 'claim', 'that', 'novel', 'coronavirus', 'was', 'never', 'detected', 'in', 'major', 'chinese', 'cities', 'of', 'beijing', 'and', 'shanghai', 'has', 'been', 'shared', 'repeatedly', 'on', 'facebook', 'twitter', 'and', 'instagram', '.']\n",
            "pred:  1 label:  1 ['facility', 'is', 'part', 'of', 'vaccines', 'pillar', 'of', 'access', 'to', 'tools', 'accelerator', 'which', 'is', 'co-led', 'by', 'and', 'who', '.', '.', '.']\n",
            "pred:  0 label:  0 ['image', 'has', 'been', 'shared', 'thousands', 'of', 'times', 'in', 'multiple', 'posts', 'on', 'facebook', 'alongside', 'a', 'claim', 'that', 'it', 'shows', 'a', 'medicine', 'created', 'by', 'u', '.', 's', '.', 'scientists', 'that', 'can', 'cure', 'novel', 'coronavirus', '.', '_']\n",
            "pred:  1 label:  1 ['rt', '.', 'is', 'providing', 'million', 'for', 'testing', 'in', 'rural', 'communities', '.', 'further', 'expanding', 'testing', 'capacity', 'is', 'a']\n",
            "pred:  1 label:  1 ['important', 'to', 'have', 'in-person', 'newborn', 'visits', 'to', 'assess', 'health', 'of', 'mom', 'and', 'baby', '.', 'talk', 'with', 'parents', 'about', 'increased', 'stress', 'of', 'having', 'a', 'new', 'baby', 'during', 'pandemic', '.', 'know', 'symptoms', 'of', 'in', 'children', '.', '.', '.']\n",
            "pred:  0 label:  0 ['quinine', 'found', 'in', 'tonic', 'water', 'along', 'with', 'of', 'zinc', 'daily', 'will', 'kill', '.', 'is', 'synthetic', 'version', 'of', 'quinine', '.', 'drink', 'some', 'fresh', 'squeezed', 'lemon', 'juice', 'with', 'your', 'tonic', 'water', '.', 'be', 'golden', '.']\n",
            "pred:  1 label:  1 ['australia', 'appears', 'to', 'be', 'heading', 'for', 'its', 'lowest', 'daily', 'increase', 'in', 'cases', 'in', 'three', 'months', 'as', 'a', 'state', 'leader', 'said', 'infection', 'spikes', 'in', 'europe', 'served', 'as', 'a', 'warning', 'about', 'dangers', 'of', 'exiting', 'lockdown', 'too', 'soon', '.']\n",
            "pred:  0 label:  0 ['lockdown', 'in', 'south', 'africa', 'has', 'collapsed', '.', 'guys', 'are', 'out', 'on', 'streets', 'looting', 'shops', 'because', 'of', 'no', 'employment', 'and', 'no', 'money', 'to', 'buy', 'food', 'even', 'police', 'cannot', 'control', 'crowds', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'we', 'all', 'have', 'a', 'role', 'to', 'play', 'in', 'preventing', 'spread', 'of', '.', '.']\n",
            "pred:  1 label:  1 ['over', 'next', 'week', 'we', 'are', 'projecting', 'arrivals', 'and', 'departures', 'from', 'managed', 'isolation', '.', 'this', 'gives', 'a', 'net', 'reduction', 'of', '344', 'people', 'in', 'managed', 'isolation', 'over', 'those', '7', 'days', '.']\n",
            "pred:  1 label:  1 ['arizona', 'has', '1300', 'new', 'cases', 'each', 'day', '.', '.']\n",
            "pred:  0 label:  0 ['poland', 'stands', 'up', 'against', 'compulsory', 'vaccination', '.']\n",
            "pred:  0 label:  0 ['people', 'have', 'to', 'be', 'in', 'fear', 'of', 'their', 'lives', 'of', 'this']\n",
            "pred:  0 label:  0 ['says', 'asian', ',', 'hong', 'kong', ',', 'swine', 'and', 'bird', 'flus', 'each', '?', '?', '?', 'killed', 'more', 'people', 'than', 'coronavirus', '.', '?', '?']\n",
            "pred:  1 label:  1 ['mortality', 'rate', 'in', 'black', 'americans', 'is', '4x', 'that', 'of', 'white', 'double', 'what', 'cdc', 'reports', '.', '.', '.']\n",
            "pred:  1 label:  1 ['is', 'it', 'difficult', 'to', 'stay', 'on', 'top', 'of', 'latest', 'research', '?', 'cdc', 'has', 'launched', 'science', 'a', 'series', 'of', 'summaries', 'of', 'new', 'studies', 'on', 'many', 'topics', '.', 'access', 'summaries', '.', '.', '.']\n",
            "pred:  1 label:  1 ['304', 'new', 'cases', 'of', ';', 'confirmed', 'discharged', '910', 'deaths', '.']\n",
            "pred:  1 label:  1 ['update', 'there', 'are', 'two', 'new', 'cases', 'of', 'to', 'report', 'in', 'managed', 'isolation', 'facilities', 'in', 'new', 'zealand', 'today', '.', 'it', 'has', 'been', '96', 'days', 'since', 'last', 'case', 'of', 'was', 'acquired', 'locally', 'from', 'an', 'unknown', 'source', '.']\n",
            "pred:  1 label:  1 ['our', 'discharges', 'today', 'include', '100', 'community', 'recoveries', 'managed', 'in', 'oyo', 'state', 'in', 'line', 'with', 'new', 'case', 'management', 'guidelines', '.', 'a', 'breakdown', 'of', 'cases', 'by', 'state', 'can', 'be', 'found', 'via', '.', '.']\n",
            "pred:  0 label:  0 ['one', 'can', 'get', 'free', 'masks', 'from', 'government', 'to', 'fight', 'coronavirus', 'by', 'filling', 'this', 'form', 'in', 'web', 'link', '.']\n",
            "pred:  0 label:  0 ['cdc', 'joins', 'forces', 'with', 'mosquito', 'joe', 'for', 'targeted', 'virus', 'control', '.']\n",
            "pred:  1 label:  1 ['yup', '.', 'and', 'yet', 'we', 'pull', 'negative', 'numbers', 'or', 'totals', 'go', 'down', '.', 'still', 'holding', 'out', 'hope', 'all', 'numbers', 'will', 'show', 'up', '.']\n",
            "pred:  1 label:  1 ['this', 'list', 'considered', 'metrics', 'including', 'social', 'distancing', 'and', 'surge', 'capacity', '.', '.', '.']\n",
            "pred:  0 label:  0 ['world', 'bank', 'documented', 'existence', 'of', 'test', 'kits', 'since', '2017', '.']\n",
            "pred:  0 label:  0 ['boiled', 'garlic', 'water', 'for', 'treating', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['staff', 'of', 'gold', 'coast', 'hospital', 'in', 'australia', 'has', 'gifted', 'a', 'volleyball', 'to', 'tom', 'hanks', 'in', 'quarantine', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '25', '.', '4', 'million', 'tests', 'up', 'from', 'yesterday', '.', '7-day', 'average', 'is', 'still', 'below', 'minimum', 'daily', 'tests', 'recommended', 'by', '.', 'for', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['currently', 'most', 'cases', 'of', 'in', 'us', 'are', 'in', 'california', 'and', 'washington', 'state', '.', 'however', 'many', 'other', 'communities', 'are', 'also', 'dealing', 'with', 'cases', 'of', '.', 'see', 'cdc', 'recommendations', 'for', 'preventing', 'spread', 'of', 'in', 'communities', '.', '.', '.']\n",
            "pred:  1 label:  1 ['michigan', 'continues', 'to', 'be', 'a', 'real', 'trouble', 'spot', 'and', 'now', 'has', 'third', 'highest', 'number', 'of', 'confirmed', 'cases', 'ny', 'and', 'crossing', '10000', 'total', 'today', '.', 'know', 'nearly', 'enough', 'about', 'their', 'testing', 'regime', 'in', 'state', 'either', '.', ')', '.']\n",
            "pred:  1 label:  1 ['arizona', 'put', 'out', 'a', 'somewhat', 'encouraging', 'set', 'of', 'numbers', '.', 'unclear', 'if', 'this', 'is', 'beginning', 'of', 'a', 'real', 'downward', 'trend', 'or', 'just', 'some', 'bumps', 'on', 'an', 'upward', 'trajectory', '.', '.']\n",
            "pred:  1 label:  1 ['children', 'under', '18', 'of', 'patients', 'were', 'most', 'likely', 'to', 'get', 'infected', 'from', 'their', 'parent', 'finds', 'a', 'new', 'study', '.', 'learn', '.', '.', '.']\n",
            "pred:  0 label:  0 ['russian', 'vaccine', 'to', 'be', 'tested', 'on', 'salisbury', 'door', 'handles', '.', '.']\n",
            "pred:  0 label:  0 ['while', 'cases', 'are', 'still', 'rising', 'there', 'should', 'be', 'a', 'study', 'on', 'patients', 'who', 'consumed', 'distributed', 'by', 'state', 'health', 'department', 'in', 'gujarat', '.', 'a', 'probe', 'should', 'be', 'done', 'if', 'any', 'of', 'them', 'were', 'diagnosed', 'with', '19', 'later', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'updates', 'new', 'cases', 'of', 'cases', 'rises', 'to', '44', 'as', '@', '8', '.', 'on', '26th']\n",
            "pred:  1 label:  1 ['these', 'high', 'levels', 'of', 'recovery', 'have', 'resulted', 'in', 'a', 'increase', 'in', 'number', 'of', 'recovered', 'cases', 'in', 'past', '30', 'days', '.', '.']\n",
            "pred:  0 label:  0 ['making', 'me', 'look', 'trump', 'blasts', 'americans', 'for', 'dying', 'from', '.']\n",
            "pred:  1 label:  1 ['they', 'provide', 'all', 'numbers', 'for', 'public', 'labs', 'but', 'not', 'negatives', 'from', 'commercial', 'tests', '.', 'like', 'close', 'but', 'no', 'cigar', '.']\n",
            "pred:  1 label:  1 ['there', 'are', 'now', '63', 'people', 'linked', 'to', 'community', 'cluster', 'who', 'remain', 'in', 'auckland', 'quarantine', 'facility', 'which', 'includes', '47', 'people', 'who', 'have', 'tested', 'positive', 'for', 'and', 'their', 'household', 'contacts', '.']\n",
            "pred:  1 label:  1 ['also', 'correction', 'on', 'first', 'tracked', '11', '.', '5', 'million', 'tests', 'not', '1', '.', '5', 'a', 'bonus', 'cumulative', 'chart', '.', '.']\n",
            "pred:  0 label:  0 ['dislike', 'of', 'u', '.', 's', '.', 'president', 'donald', 'trump', 'supersedes', 'their', 'desire', 'to', 'treat', 'coronavirus', 'with', 'donated', 'plasma', '.']\n",
            "pred:  1 label:  1 ['.', '.', '.', 'of', 'cases', '.', 'as', 'on', '7', '.', '9', '.', '2020', '.', 'moreover', 'goi', 'has', 'fast-tracked', 'approval', 'of', 'testing', 'kits', 'and', 'clinical', 'trials', 'of', 'vaccine', ';', 'medicines', 'for', '.', 'meanwhile', 'people', 'can', 'contribute', 'towards', 'controlling', 'by', '.', '.', '.']\n",
            "pred:  1 label:  1 ['maintains', 'its', 'high', 'recovery', 'total', 'recoveries', 'cross', '27', 'lakh', 'recovered', 'patients', 'are', 'more', 'than', '3', '.', '5', 'times', 'active', 'cases', '.', '.']\n",
            "pred:  0 label:  0 ['trump', 'announced', 'that', 'roche', 'medical', 'company', 'will', 'launch', 'vaccine', 'next', 'sunday', ',', 'and', 'millions', 'of', 'doses', 'are', 'ready', 'from', 'it', '!!!', 'end', 'of', 'play']\n",
            "pred:  1 label:  1 ['nationally', '7-day', 'average', 'in', 'deaths', 'has', 'begun', 'to', 'rise', 'after', 'an', 'extended', 'decline', '.', 'last', 'three', 'days', 'were', 'highest', 'numbers', 'since', 'early', 'june', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'facebook', 'post', 'that', 'argues', 'that', 'florida', 'compares', 'favorably', 'with', 'new', 'jersey', 'and', 'new', 'york', 'in', 'rate', 'of', 'deaths', 'has', 'a', 'point', '.', 'but', 'timing', 'is', 'a', 'big', 'reason', 'why', 'post', 'is', 'misleading', '.', '.']\n",
            "pred:  1 label:  1 ['a', 'study', 'from', 'journal', 'of', 'medical', 'virology', 'shows', 'that', 'it', 'takes', 'more', 'than', '2', 'weeks', 'for', 'patients', 'to', 'clear', 'viral', 'rna', '.', 'with', 'that', 'said', 'study', 'also', 'found', 'that', 'people', 'with', 'milder', 'symptoms', 'tend', 'to', 'take', 'longer', 'to', 'clear', 'virus', '.', 'read', 'more', '.']\n",
            "pred:  0 label:  0 ['humans', 'to', 'be', 'allowed', 'out', 'of', 'temporary', 'lockdown', 'to', 'see', 'animals', 'in', 'permanent', 'lockdown', '.', '.']\n",
            "pred:  1 label:  1 ['one', 'crucial', 'addendum', 'to', 'this', 'piece', 'by', 'most', 'important', 'piece', 'of', 'missing', 'data', 'would', 'be', 'breakouts', 'by', 'race', 'for', 'deaths', 'not', 'just', 'cases', '.', 'not', 'a', 'single', 'state', 'reports', 'that', 'yet', '.', '.']\n",
            "pred:  1 label:  1 ['a', 'number', 'of', 'vaccines', 'are', 'now', 'in', 'phase', 'three', 'clinical', 'trials', 'and', 'we', 'all', 'hope', 'to', 'have', 'a', 'number', 'of', 'effective', 'vaccines', 'that', 'can', 'help', 'prevent', 'people', 'from', 'infection', '.', 'however', 'no', 'silver', 'bullet', 'at', 'moment', 'and', 'there', 'might', 'never']\n",
            "pred:  0 label:  0 ['uss', 'bonhomme', 'richard', 'boasts', 'zero', 'cases', '.']\n",
            "pred:  0 label:  0 ['coping', 'with', 'coronavirus', 'boredom', '.']\n",
            "pred:  0 label:  0 ['once', 'over', 'going', 'to', 'new', 'and', 'getting', 'a', 'visible', 'crocodile', 'skin', 'tattoo', 'so', 'my', 'mom', 'will', 'ever', 'know', '.', '.']\n",
            "pred:  1 label:  1 ['this', 'case', 'brings', 'our', 'total', 'number', 'of', 'confirmed', 'cases', 'to', '1193', 'which', 'is', 'number', 'we', 'report', 'to', 'world', 'health', 'organization', '.', 'there', 'is', 'no', 'one', 'in', 'new', 'zealand', 'receiving', 'care', 'for', '.']\n",
            "pred:  1 label:  1 ['.', '.', '.', 'which', 'an', 'individual', 'with', 'is', 'infectious', 'is', 'uncertain', '.', 'as', 'per', 'current', 'evidence', 'period', 'of', 'infectivity', 'starts', '2', 'days', 'prior', 'to', 'onset', 'of', 'symptoms', 'and', 'lasts', 'up', 'to', '8', '.']\n",
            "pred:  1 label:  1 ['there', 'are', 'four', 'patients', 'with', 'in', 'middlemore', 'hospital', '.', 'two', 'are', 'stable', 'and', 'each', 'of', 'these', 'is', 'in', 'isolation', 'on', 'a', 'ward', '.', 'two', 'are', 'in', 'icu', 'and', 'are', 'in', 'critical', 'condition', '.', 'these', 'are', 'same', 'four', 'patients', 'previously', 'reported', 'and', 'are', 'all', 'part', 'of', 'community', 'cluster', '.']\n",
            "pred:  1 label:  1 ['vaccine', 'is', 'ready', 'to', 'ship', 'as', 'soon', 'as', 'fda', 'approves', 'it', '.', '.']\n",
            "pred:  0 label:  0 ['donald', 'trump', 'calls', 'coronavirus', 'virus', '.', '’', 'health', 'experts', 'say', '.', '.']\n",
            "pred:  0 label:  0 ['2005', 'study', 'found', 'that', 'chloroquine', 'had', 'inhibitory', 'effects', 'on', 'animal', 'cells', 'infected', 'with', 'sars', 'not', '.', 'drug', 'given', 'to', 'actual', 'human', 'patients', 'and', 'authors', 'wrote', 'that', 'more', 'research', 'was', 'needed', '.', '.']\n",
            "pred:  0 label:  0 ['before', 'corona', 'we', 'were', 'al', 'bouncing', 'get', 'back', 'there', 'again']\n",
            "pred:  1 label:  1 ['is', 'when', 'everyone', 'has', 'opportunity', 'to', 'be', 'as', 'healthy', 'as', 'possible', '.', 'learn', 'about', 'role', 'of', 'culture', 'in', 'your', 'efforts', 'to', 'stop', '.', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'total', 'confirmed', 'cases', 'cases', 'samples', 'tested']\n",
            "pred:  0 label:  0 ['in', 'italy', 'cure', 'for', 'coronavirus', 'is', 'finally']\n",
            "pred:  0 label:  0 ['a', 'nurse', 'working', 'in', 'us', 'informs', 'ivorian', 'citizens', 'that', 'their', 'government', 'is', 'misleading', 'them', 'disguising', 'vaccines', 'centers', 'into', 'testing', 'centers', 'against', '.']\n",
            "pred:  0 label:  0 ['who', 'says', 'that', 'has', 'become', 'endemic', '.']\n",
            "pred:  1 label:  1 ['sadly', 'three', 'people', 'are', 'in', 'hospital', 'with', '–', 'one', 'each', 'at', 'auckland', 'city', 'middlemore', 'and', 'north', 'shore', 'hospitals', '.', 'all', 'three', 'patients', 'are', 'in', 'isolation', 'on', 'a', 'general', 'ward', '.']\n",
            "pred:  1 label:  1 ['health', 'minister', 'edward', 'says', 'a', 'two', 'week', 'national', 'lockdown', 'is', 'not', 'something', '.', 'get', 'latest', 'on', 'this', 'story', '.', '.']\n",
            "pred:  0 label:  0 ['report', 'says', '5', ',', '000', 'students', 'taking', 'got', '.']\n",
            "pred:  1 label:  1 ['teenager', 'has', 'been', 'interviewed', 'thoroughly', 'to', 'ensure', 'we', 'can', 'identify', 'any', 'other', 'potential', 'contacts', 'and', 'take', 'appropriate', 'action', '.', 'everybody', 'coming', 'in', 'on', 'that', 'flight', 'from', 'melbourne', 'to', 'auckland', 'is', 'going', 'into', 'managed', 'isolation', 'where', 'they', 'are', 'all', 'being', 'tested', '.']\n",
            "pred:  0 label:  0 ['video', 'shows', 'president', 'donald', 'trump', 'saying', 'is', 'democrats', '?', '?', '?', '?', '?', '?', 'new', 'hoax', '.', '?', '?']\n",
            "pred:  1 label:  1 ['rt', 'latest', 'report', 'from', 'shows', 'that', 'parts', 'of', 'u', '.', 's', '.', 'are', 'seeing', 'increases', 'in', 'some', 'indicators', 'used', 'to']\n",
            "pred:  0 label:  0 ['need', 'to', 'open', 'up', 'economy', 'and', 'get', 'back', 'to', 'work', ',', '”', 'says', '.', '.']\n",
            "pred:  1 label:  1 ['whole', 'crew', 'led', 'by', 'have', 'been', 'on', 'since', 'beginning', 'and', 'have', 'released', 'important', 'arguments', 'about', 'how', 'to', 'frame', 'disproportionate', 'impacts', 'of', 'disease', 'on', 'black', 'communities', '.', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'act', 'now', 'is', 'a', 'tool', 'to', 'help', 'leaders', 'and', 'communities', 'understand', 'how', 'pandemic', 'will', 'affect', 'regions', 'across']\n",
            "pred:  0 label:  0 ['7', 'kids', 'died', 'in', 'senegal', 'right', 'after', 'receiving', 'vaccine', 'for', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['that', 'indian', 'prime', 'minister', 'modi', 'said', 'one', 'crore', 'positive', 'patients', 'have', 'been', 'treated', 'for', 'free', '.', '_']\n",
            "pred:  0 label:  0 ['gov', '.', 'newsom', 'puts', 'coronavirus', 'bounty', 'on', 'straying', 'senior', 'citizens', '.']\n",
            "pred:  0 label:  0 ['a', 'super', 'was', 'created', 'in', 'a', 'chinese', 'biosecurity', 'lab', 'and', 'has', 'infected', 'more', 'than', '100', ',', '000', 'people', 'in', 'wuhan', '.']\n",
            "pred:  1 label:  1 ['yesterday', 'our', 'laboratories', 'processed', '1692', 'tests', 'of', 'which', '1259', 'swabs', 'were', 'taken', 'in', 'community', 'and', '433', 'swabs', 'were', 'taken', 'in', 'managed', 'isolation', 'or', 'quarantine', 'facilities', '.']\n",
            "pred:  0 label:  0 ['an', 'mla', 'hafeez', 'khan', 'forced', 'a', 'nurse', 'to', 'touch', 'feet', 'of', 'a', 'muslim', 'priest', 'because', 'she', 'criticized', 'jamat', 'for', 'spreading', 'coronavirus', 'in', 'india', '.']\n",
            "pred:  0 label:  0 ['donald', 'trump', 'ignored', 'experts', ',', 'downplayed', 'threat', 'posed', ',', 'and', 'failed', 'to', 'take', 'action', 'necessary', 'to', 'contain', 'it', '—', 'and', 'now', 'in', 'many', 'states', ',', 'cases', 'are', 'on', 'rise', 'again', '.', 'this', 'president', 'has', 'failed', 'to', 'lead', 'our', 'nation', '.', '.']\n",
            "pred:  0 label:  0 ['facebook', 'user', 'suggests', 'that', 'confirmed', 'cases', 'are', 'suspected', 'cases', 'of', 'whereas', 'those', 'declared', 'positive', 'are', 'people', 'who', 'actually', 'have', 'novel', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['chinese', 'medical', 'supplies', 'to', 'ghana', 'contaminated', 'with', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['as', 'of', 'today', 'acc', 'to', '2', '.', 'of', 'total', 'active', 'cases', 'are', 'currently', 'in', 'icu', '.', '.']\n",
            "pred:  1 label:  1 ['there', 'are', '70', 'people', 'linked', 'to', 'community', 'cluster', 'who', 'remain', 'in', 'auckland', 'quarantine', 'facility', 'which', 'includes', '52', 'people', 'who', 'have', 'tested', 'positive', 'for', 'and', 'their', 'household', 'contacts', '.']\n",
            "pred:  0 label:  0 ['tantrum', 'diplomacy', 'is', 'eroding', 'u', '.', 's', '.', 'credibility', 'on', 'international', 'stage', '.', 'withdrawing', 'from', 'during', 'a', 'global', 'pandemic', 'is', 'a', 'recipe', 'for', 'disaster', 'and', 'increases', 'cost', 'of', 'american', 'response', '.', 'we', 'must', 'instead', 'rally', 'world', 'to', 'confront', '.', '.']\n",
            "pred:  1 label:  1 ['issues', 'guidelines', 'for', 'enforcement', 'of', 'lockdown', 'in', 'containment', 'zones', 'till', '30th', 'september', '2020', 'persons', 'advised', 'to', 'stay', 'home', 'via', '.']\n",
            "pred:  1 label:  1 ['fda', 'chief', 'hahn', 'comments', 'about', 'convalescent', 'plasma', 'for', 'patients', 'stirred', 'controversy', 'and', 'a', 'shake', 'up', 'at', 'agency', '.', '.']\n",
            "pred:  0 label:  0 ['myanmar', 'president', 'and', 'state', 'counselor', 'are', 'under', '14', 'days', 'quarantine', 'after', 'meeting', 'with', 'a', 'patient', 'from', 'switzerland', '.']\n",
            "pred:  0 label:  0 ['nurses', 'submitting', 'coronavirus', 'tests', 'say', 'they', 'all', 'are', 'coming', 'back', 'positive', '.']\n",
            "pred:  0 label:  0 ['group', 'hanging', 'out', 'in', 'park', 'presumably', 'celebrating', 'blood', 'on', 'their', 'hands', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'how', 'are', 'vaccine', 'trials', 'performed', '?', 'this', 'short', 'video', '.']\n",
            "pred:  1 label:  1 ['pregnant', 'women', 'with', 'may', 'be', 'at', 'increased', 'risk', 'for', 'severe', 'illness', 'compared', 'with', 'women', '.', 'case', 'data', 'are', 'now', 'available', 'including', 'age', 'race/ethnicity', 'and', 'need', 'for', 'critical', 'care', '.', 'see', 'this', '.', '.', '.']\n",
            "pred:  0 label:  0 ['japanese', 'nobel', 'laureate', 'says', 'coronavirus', 'is', 'not', 'naturally', 'occurring', '.']\n",
            "pred:  1 label:  1 ['a', 'spanish', 'medic', 'has', 'warned', 'that', 'unless', 'britons', 'follow', 'new', 'lockdown', 'rules', 'and', 'wear', 'masks', 'could', 'be', 'very', 'very', '.']\n",
            "pred:  0 label:  0 ['stop', 'until', 'coronavirus', 'has', 'spread', 'to', 'hong', 'xi', 'jinping', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '19', '.', '8', 'million', 'tests', 'up', 'from', 'yesterday', 'close', 'to', 'all-time', 'high', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'for', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'crosses', 'a', 'crucial', 'milestone', 'in', 'fight', 'against', 'more', 'than', '10', 'lakh', 'people', 'in', 'a', 'day']\n",
            "pred:  1 label:  1 ['at', 'height', 'of', 'pandemic', 'in', 'new', 'york', 'city', 'first', 'responders', 'faced', 'nearly', 'triple', 'number', 'of', 'calls', 'for', 'cardiac', 'arrest', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'this', 'lady', 'here', 'applied', 'sanitizer', 'to', 'her', '&', 'went', 'to', 'kitchen', 'to', 'cook', '.', 'moment', 'she', 'turned', 'on', 'gas', 'stove', ',', 'her', 'hands', 'caught', 'fire', 'due', 'to', 'alcohol', 'contained', 'in', 'sanitizer', '.', '?', '?']\n",
            "pred:  1 label:  1 ['rt', 'for', 'first', 'time', 'number', 'of', 'active', 'cases', 'has', 'reduced', 'by', 'last', '24', '.', 'more', 'than', 'of', 'total']\n",
            "pred:  1 label:  1 ['states', 'reported', 'new', 'cases', 'in', 'line', 'with', 'slow', 'drift', 'downward', '.', '.']\n",
            "pred:  0 label:  0 ['when', 'shri', 'became', '19', 'positive', ',', 'were', 'organizing', 'funeral', '.']\n",
            "pred:  0 label:  0 ['avoid', 'cold', 'drinks', 'and', 'sweets', 'for', '90', 'days', 'to', 'prevent', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['washington', 'state', 'doh', 'noted', 'that', 'they', 'will', 'no', 'longer', 'report', 'deaths', 'over', 'weekend', 'starting', 'today', '.', 'weekend', 'deaths', 'in', 'state', 'will', 'be', 'included', 'in', 'monday', 'and', 'tuesday', 'counts', '.']\n",
            "pred:  1 label:  1 ['national', 'case', 'fatality', 'rate', 'for', 'further', 'dips', 'to', '2', '.', '.', 'india', 'has', 'registered', 'one', 'of', 'lowest', 'deaths', 'per', 'million', 'at', '30', 'as', 'compared', 'to', 'global', 'average', 'of', '91', 'deaths', 'per', 'million', '.', '.']\n",
            "pred:  0 label:  0 ['latest', 'polls', 'show', 'fifty-two', 'percent', 'of', 'britons', 'will', 'refuse', 'to', 'take', 'coronavirus', 'vaccine', '.', '.']\n",
            "pred:  1 label:  1 ['doing', 'over', 'tests', 'a', 'day', 'health', 'secretary', 'matt', 'hancock', 'insists', 'over', 'tests', 'are', 'being', 'carried', 'out', 'each', 'day', 'across', 'uk', 'in', 'lighthouse', 'labs', 'and', 'nhs', 'read', 'more', '.', '.']\n",
            "pred:  1 label:  1 ['everyone', 'who', 'has', 'left', 'managed', 'isolation', 'since', '9th', 'of', 'june', 'have', 'been', 'tested', 'for', 'at', 'facility', 'or', 'subsequently', '.', 'for', 'two', 'women', 'who', 'arrived', 'back', 'from', 'uk', 'and', 'were', 'in', 'novotel', 'ellerslie', 'everybody', 'at', 'hotel', 'at', 'that', 'time', 'was', 'tested', 'and', 'returned', 'negative', 'tests', '.']\n",
            "pred:  0 label:  0 ['we', 'will', 'all', 'come', 'out', 'stronger', 'from', 'this', '.', 'just', '.']\n",
            "pred:  1 label:  1 ['rt', 'total', 'confirmed', 'cases', 'cases', 'samples', 'tested']\n",
            "pred:  1 label:  1 ['spread', 'of', 'coronavirus', 'through', 'fans', 'and', 'air', 'conditioning', 'is', 'this', 'clinical', 'trend', '.', '.', '.']\n",
            "pred:  1 label:  1 ['three', 'stories', 'to', 'read', 'this', 'records', 'new', 'cases', '-', 'highest', 'daily', 'figure', 'in', 'four', 'months', '.', 'johnson', 'reveals', 'new', 'rules', '.', 'bans', 'household', 'visits', '.', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '13', '.', '4', 'million', 'tests', 'up', 'from', 'yesterday', 'substantially', 'below', '7-day', 'average', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'for', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['here', 'are', 'states', 'reporting', 'over', '400', 'cases', 'today', '.', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'states', 'reported', 'tests', 'and', 'cases', '.', 'death', 'toll', 'today', 'is', '348', 'lowest', 'number', 'since', 'july', '12', '.', '.']\n",
            "pred:  1 label:  1 ['in', 'a', 'community', 'setting', 'remember', '3', 'wash', 'your', 'hands', '.', 'watch', 'your', 'distance', '6', 'feet', '.', '.', 'these', 'habits', 'can', 'help', 'you', 'protect', 'yourself', 'and', 'others', 'from', 'spreading', '.', 'learn', '.', '.', '.']\n",
            "pred:  0 label:  0 ['trump', 'says', 'coronavirus', 'task', 'force', 'will', 'keep', 'working', 'with', 'a', 'focus', 'on', 'vaccines', 'and', 'reopening', '.']\n",
            "pred:  0 label:  0 ['hospitals', 'allow', 'parents', 'to', 'visit', 'children', 'with', 'in', 'hospital', '.']\n",
            "pred:  0 label:  0 ['baba', 'ramdev', 'claims', 'that', 'can', 'be', 'self', 'diagnosed', 'by', 'controlling', 'breathing', 'and', 'that', 'mustard', 'oil', 'kills', 'virus', '.']\n",
            "pred:  1 label:  1 ['one', 'last', 'saturday', 'note', 'on', 'partially', 'testing', 'plateau', '.', 'avg', 'number', 'of', 'completed', 'tests', 'over', 'last', '7', 'avg', 'number', 'of', 'completed', 'tests', 'over', '7', 'days', 'before', 'yet', 'so', 'many', 'talking', 'about', 'scaling', 'up', 'testing', '.', '.', '.']\n",
            "pred:  0 label:  0 ['indian', 'economist', 'raghuram', 'rajan', 'chaired', 'imf', 'webinar', 'on', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['64', 'new', 'cases', 'of', 'reported', ';', 'as', 'at', '27th', '1337', 'confirmed', 'cases', 'of', 'reported', 'in', 'nigeria', '.', '255', '40', '.']\n",
            "pred:  0 label:  0 ['joe', 'biden', 'says', 'is', 'deadliest', 'threat', 'cops', 'face', '.', 'numbers', 'back', 'him', 'up', '.', '.', '.']\n",
            "pred:  1 label:  1 ['india', 'registers', 'recoveries', 'of', 'in', 'last', '24', 'hours', 'whereas', 'new', 'cases', 'were', 'reported', 'in', 'last', '24', 'hours', '819', 'deaths', 'reported', '.']\n",
            "pred:  0 label:  0 ['video', 'of', 'dead', 'bodies', 'being', 'piled', 'up', 'in', 'italy', 'due', 'to', '.']\n",
            "pred:  0 label:  0 ['sir', 'ganga', 'ram', 'hospital', ',', 'delhi', 'issue', 'prescription', '.']\n",
            "pred:  1 label:  1 ['rt', 'due', 'to', 'large', 'amount', 'of', 'speculation', 'regarding', 'this', 'is', 'a', 'reminder', 'to', 'rely', 'on', 'official', 'sources', 'for', 'accurate']\n",
            "pred:  1 label:  1 ['overall', 'mortality', 'of', 'patients', 'with', 'in', 'icus', 'has', 'dropped', 'from', '~', 'in', 'march', 'to', '~', 'in', 'may', 'a', 'new', 'analysis', 'finds', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'great-great-grandmother', 'is', 'arrested', 'for', 'making', 'coronavirus', 'masks', 'out', 'of', 'her', 'granny', '.']\n",
            "pred:  0 label:  0 ['corona', 'syrup', 'is', 'better', 'than', 'turmeric', '-1', 'spoon', ',', 'clove', '-2', ',', 'lemon', 'with', 'skin', '-1', ',', 'skin', 'peeled', 'ginger', ',', 'plus', 'one', 'litre', 'water', '(', 'boil', 'it', 'till', 'it', 'becomes', 'to', 'make', 'syrup', 'note', ':', 'suggestion', 'from', 'an', 'doctor', 'who', 'got', 'cured', 'from', 'coronavirus', 'within', '24', 'hours', '.']\n",
            "pred:  1 label:  1 ['this', 'is', 'just', 'another', 'indication', 'that', 'while', 'it', 'broke', 'out', 'on', 'east', 'and', 'west', 'coasts', 'will', 'not', 'remain', 'a', 'coastal', 'phenomenon', 'as', 'michigan', 'louisiana', 'illinois', 'indiana', 'colorado', 'and', 'other', 'states', 'experience', 'intense', 'outbreaks', '.']\n",
            "pred:  1 label:  1 ['yup', '.', 'prospective', 'data', 'hole', 'really', 'opening', '.', 'very', 'concerned', 'about', 'it', 'but', 'seen', 'states', 'like', 'tn', 'sometimes', 'move', 'towards', 'openness', '.']\n",
            "pred:  1 label:  1 ['centre', 'exhorts', 'north', 'eastern', 'states', 'to', 'actively', 'break', 'chain', 'of', 'transmission', '.', 'these', '8', 'north', 'eastern', 'states', 'together', 'account', 'for', 'less', 'than', 'of', 'total', 'active', 'cases', 'in', 'country', '.', '.', '.']\n",
            "pred:  0 label:  0 ['an', 'alleged', 'bot', 'system', 'on', 'twitter', 'controlled', 'by', 'right', 'wing', 'of', 'spain', 'uncovered', 'by', 'mistake', '.']\n",
            "pred:  1 label:  1 ['healthcare', 'workers', 'in', 'tx', 'az', 'and', 'fl', 'hot', 'spots', 'are', 'more', 'prepared', 'for', 'coronavirus', 'than', 'ny', 'was', 'but', 'still', 'fighting', 'anxiety', 'and', 'exhaustion', '.', '.']\n",
            "pred:  0 label:  0 ['trump', 'says', 'he', 'would', 'mobilize', 'military', 'to', 'distribute', 'coronavirus', 'vaccine', 'when', 'ready', '.']\n",
            "pred:  0 label:  0 ['a', 'lot', 'more', 'to', 'be', 'learned', 'about', 'but', 'we', 'found', 'no', 'evidence', 'that', 'harmless', 'for', 'of', 'people', 'it', 'touches', 'as', 'trump', 'falsely', 'claimed', '.', '.']\n",
            "pred:  0 label:  0 ['viral', 'whatsapp', 'message', 'advises', 'against', 'purchasing', 'used', 'clothing', 'as', 'they', 'expose', 'buyers', 'to', '.']\n",
            "pred:  1 label:  1 ['we', 'just', 'updated', 'race', 'and', 'ethnicity', 'data', 'in', 'racial', 'data', 'tracker', 'and', 'happy', 'to', 'report', 'progress', '.', 'only', 'four', 'states', 'do', 'not', 'report', 'some', 'kind', 'of', 'race/ethnicity', 'data', '.', 'they', 'north', 'dakota', 'nebraska', 'nevada', 'and', 'south', 'dakota', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'tested', 'for', '2019', 'novel', 'coronavirus', 'at', 'makati', 'medical', 'center', '.']\n",
            "pred:  0 label:  0 ['claim', 'that', 'dr', '.', 'fauci', 'said', 'every', 'american', 'should', 'be', 'microchipped', '.']\n",
            "pred:  1 label:  1 ['our', 'national', 'contact', 'tracing', 'centre', 'has', '140', 'people', 'on', 'standby', 'to', 'support', 'auckland', 'regional', 'public', 'health', '.', 'healthline', 'has', 'been', 'very', 'busy', 'as', 'we', 'anticipated', 'with', 'high', 'volumes', 'of', 'calls', '.', 'they', 'have', 'an', 'extra', '60', 'staff', 'already', 'and', 'we', 'will', 'be', 'increasing', 'that', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '24', '.', '9', 'million', 'tests', 'up', 'from', 'yesterday', 'as', '7-day', 'average', 'approaches', 'minimum', 'daily', 'tests', 'recommended', 'by', '.', 'for', 'details', '.', '.']\n",
            "pred:  0 label:  0 ['gates', 'foundation', 'stands', 'to', 'make', 'nearly', '?', '?', '31', '.', '5', 'billion', 'on', 'a', 'coronavirus', 'vaccine', 'in', 'u', '.', 'k', '.']\n",
            "pred:  0 label:  0 ['dr', 'vk', 'srinivas', 'vice', 'president', 'of', 'bharath', 'bio', 'tech', 'he', 'taking', 'first', 'injection', 'of', '19', 'they', 'have', 'confidence', 'of', 'their', 'product']\n",
            "pred:  1 label:  1 ['case', 'fatality', 'rate', 'continues', 'to', 'slide', 'improves', 'to', '2', '.', '.', 'active', 'cases', 'as', 'a', 'percentage', 'of', 'total', 'cases', 'have', 'seen', 'a', 'significant', 'drop', 'from', '34', '.', 'on', '24th', 'july', '2020', 'to', '30', '.', 'as', 'on', 'today', '.', '.']\n",
            "pred:  1 label:  1 ['of', 'taste', 'and', 'smell', 'is', 'a', 'symptom', 'of', 'and', 'could', 'also', 'be', 'for', 'some', 'other', 'illnesses', '.', 'it', 'is', 'important', 'to', 'get', 'tested', 'rather', 'than', 'on', 'assumption', 'that', 'it', 'is', '~', 'on', '.']\n",
            "pred:  0 label:  0 ['a', 'message', 'claimed', 'that', 'indian', 'government', 'plans', 'to', 'ease', 'down', 'lockdown', 'restrictions', 'imposed', 'in', 'country', 'in', 'five', 'stages', '.']\n",
            "pred:  0 label:  0 ['they', 'have', 'called', 'coronavirus', 'something', 'else', '?', '”', 'asks', 'local', 'man', 'named', '.']\n",
            "pred:  1 label:  1 ['as', 'at', '8:00', 'pm', '27th', 'march', 'there', 'are', '70', 'confirmed', 'cases', '3', 'discharged', '1', 'death', 'for', 'a', 'breakdown', 'of', 'cases', 'by', 'states', 'in', 'real', 'time', 'please', 'see', '.', 'currently', ';', '44', '14', '3', '1', '3', '1', '2']\n",
            "pred:  1 label:  1 ['around', 'of', 'active', 'cases', 'are', 'concentrated', 'in', 'only', '5', 'most', 'affected', 'states', '.', 'there', 'are', '13', 'states', 'and', 'uts', 'that', 'even', 'today', 'have', 'less', 'than', '5000', 'active', 'cases', '.', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'does', 'not', 'affect', 'people', 'with', 'blood', 'type', '.']\n",
            "pred:  1 label:  1 ['western', 'europe', 'has', 're-emerged', 'as', 'a', 'global', 'hot-spot', 'for', '.', 'despite', 'containing', 'spread', 'of', 'virus', 'earlier', 'this', 'year', 'region', 'surpassed', 'u', '.', 's', '.', 'for', 'daily', 'new', 'cases', 'many', 'of', 'which', 'are', 'linked', 'to', 'returning', 'travelers', ';', 'socializing', '.', '.']\n",
            "pred:  0 label:  0 ['houses', 'of', 'worship', 'in', 'accounting', 'crisis', 'over', 'relief', 'package', '.']\n",
            "pred:  0 label:  0 ['toronto', 'catholic', 'schools', 'to', 'reopen', 'so', 'kids', 'will', 'be', 'with', 'god', 'sooner', '.']\n",
            "pred:  0 label:  0 ['tree', 'cures']\n",
            "pred:  1 label:  1 ['take', 'steps', 'to', 'keep', 'your', 'friends', 'and', 'family', 'safe', 'when', 'visiting', 'those', 'at', 'higher', 'risk', 'for', 'severe', 'illness', 'from', '.', 'learn', 'more', 'at', '.', '.', '.']\n",
            "pred:  1 label:  1 ['in', 'next', 'couple', 'weeks', 'we', 'will', 'have', 'a', 'dedicated', 'page', 'racial', 'data', 'tracker', 'that', 'will', 'allow', 'anyone', 'to', 'explore', 'these', 'dynamics', 'of', 'outbreak', 'using', 'both', 'data', 'we', 'collect', 'and', 'other', 'datasets', 'that', 'enrich', 'our', 'understanding', 'of', 'disparities', 'that', 'seeing', '.']\n",
            "pred:  1 label:  1 ['rt', 'registers', 'a', 'record', 'of', 'highest', 'single', 'day', 'recoveries', 'than', '1', 'lakh', 'patients']\n",
            "pred:  1 label:  1 ['rt', 'in', 'my', 'suo', 'moto', 'statement', 'i', 'informed', 'members', 'that', 'thro', '13', 'clinical', 'trials', 'of', 'repurposed', 'drugs', ';']\n",
            "pred:  1 label:  1 ['and', 'through', 'global', 'vaccines', 'facility', 'countries', 'that', 'represent', 'nearly', 'of', 'global', 'population', 'have', 'signed', 'up', 'or', 'expressed', 'an', 'interest', 'to', 'be', 'part', 'of', 'new']\n",
            "pred:  0 label:  0 ['a', 'video', 'shows', 'an', 'empty', 'triage', 'tent', 'outside', 'to', 'a', 'hospital', 'and', 'a', 'man', 'claiming', 'that', 'everything', 'is', 'a', 'lie', 'because', 'there', 'are', 'no', 'sick', 'people', 'in', 'tent', '.']\n",
            "pred:  0 label:  0 ['dr', '.', 'anthony', 'fauci', 'wrote', 'a', 'post', 'describing', 'dangers', 'of', 'coronavirus', 'and', 'criticizing', 'those', 'who', 'do', 'not', 'take', 'it', 'seriously', '.']\n",
            "pred:  1 label:  1 ['we', 'now', 'have', '1402', 'people', 'reported', 'as', 'having', 'recovered', 'from', '.', 'this', 'is', 'of', 'all', 'cases', '.', 'today', 'there', 'are', 'two', 'people', 'receiving', 'hospital', 'level', 'care', '–', 'one', 'each', 'in', 'middlemore', 'and', 'north', 'shore', 'hospitals', '.', 'neither', 'is', 'in', 'icu', '.', 'there', 'are', 'no', 'additional', 'deaths', 'to', 'report', '.']\n",
            "pred:  0 label:  0 ['u', '.', 's', '.', 'protocol', 'for', 'italy', 'were', 'discovered']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '16', '.', '9', 'million', 'tests', 'up', 'from', 'yesterday', 'total', 'recorded', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'for', 'details', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'first', 'vaccine', ',', 'while', 'a', 'welcome', 'tool', 'in', 'fighting', ',', 'may', 'turn', 'out', 'to', 'be', 'of', 'limited', 'use', '.', 'lesson', 'learned', 'from', 'aids', 'is', 'value', 'of', 'building', 'a', 'scientific', 'infrastructure', 'beyond', 'a', 'vaccine', '.', '.']\n",
            "pred:  0 label:  0 ['that', 'chinese', 'film', 'about', 'coronavirus', '?', '\"', 'asks', 'taxi', 'driver', '.']\n",
            "pred:  0 label:  0 ['was', 'cured', 'of', 'through', 'ayurveda', 'and', 'homeopathy', 'treatment', 'from', 'a', 'holistic', 'resort', ',', 'union', 'minister', 'of', 'state', 'for', 'ayush', 'shripad', 'naik', 'claimed', 'on', 'thursday', '.']\n",
            "pred:  0 label:  0 ['has', 'stated', 'russian', 'citizens', '2', 'options', 'stay', 'home', 'for', '15', 'days', 'or', 'in', 'jail', 'for', '5', 'years', '.', '”']\n",
            "pred:  0 label:  0 ['bing', 'liu', 'a', 'researcher', 'who', 'was', 'about', 'to', 'discover', 'a', 'vaccine', 'for', 'was', 'murdered', 'in', 'us', '.']\n",
            "pred:  1 label:  1 ['rt', 'your', 'risk', 'of', 'increases', 'closer', 'you', 'get', 'to', 'others', 'and', 'longer', 'you', 'interact', 'with', 'them', '.', 'learn', 'from']\n",
            "pred:  0 label:  0 ['photograph', 'proving', 'that', 'no', 'more', 'room', 'in', 'italian', 'hospitals', '.']\n",
            "pred:  1 label:  1 ['people', 'in', 'scotland', 'have', 'been', 'banned', 'from', 'visiting', 'other', 'households', 'indoors', 'as', 'rules', 'are', 'tightened', '.']\n",
            "pred:  0 label:  0 ['wearing', 'a', 'mask', 'during', 'physical', 'activity', 'causes', 'hypercapnia', 'syndrome', '.']\n",
            "pred:  1 label:  1 ['broke', 'down', 'why', 'daily', 'deaths', 'are', 'down', 'an', 'average', 'of', '75', 'percent', 'from', 'their', 'april', 'peak', 'even', 'with', 'cases', 'rising', '.', '.']\n",
            "pred:  1 label:  1 ['boris', 'plans', 'to', 'impose', 'a', '10pm', 'curfew', 'on', 'pubs', 'and', 'restaurants', 'have', 'been', 'described', 'as', 'a', 'for', 'hospitality', 'sector', '.']\n",
            "pred:  0 label:  0 ['marathi', 'newspaper', 'reports', 'a', 'conspiracy', 'theory', 'of', 'coronavirus', 'being', 'a', 'bioweapon', '.']\n",
            "pred:  1 label:  1 ['for', 'a', 'breakdown', 'of', 'cases', 'by', 'states', 'in', 'real', 'time', 'please', 'see', '.', 'currently', ';', '44', '12', '3', '1', '1', '1', '65', 'confirmed', 'cases']\n",
            "pred:  0 label:  0 ['millions', 'of', 'people', 'have', 'seen', 'a', 'viral', 'video', 'of', 'doctors', 'talking', 'about', 'coronavirus', 'pandemic', '.', 'all', 'of', 'physicians', 'we', 'have', 'a', 'history', 'of', 'making', 'unproven', 'conspiratorial', 'or', 'bizarre', 'medical', 'claims', '.', '.']\n",
            "pred:  0 label:  0 ['today', ',', 'special', 'military', 'helicopters', 'will', 'spray', 'pesticides', 'against', 'corona', 'virus', 'in', 'skies', 'all', 'over', 'uae', 'people', 'are', 'advised', 'to', 'stay', 'indoors', 'after', 'twelve', \"o'clock\", 'at', 'night', 'and', 'remove', 'all', 'clothes', 'which', 'are', 'outside', '.', 'pls', 'rt', 'for', 'those', 'with', 'in', 'dubai', 'and', 'uae', '.']\n",
            "pred:  1 label:  1 ['we', 'also', 'have', 'to', 'report', 'a', 'bug', 'in', 'our', 'rolling', 'average', 'lines', 'in', 'some', 'previous', 'charts', '.', 'in', 'some', 'cases', 'they', 'were', 'calculating', 'over', 'a', 'longer', 'time', 'period', 'which', 'made', 'changes', 'harder', 'to', 'see', '.', 'we', 'regret', 'error', '.']\n",
            "pred:  0 label:  0 ['who', 'has', 'warned', 'of', 'an', 'outbreak', 'of', 'nipah', 'in', 'india', 'than', '.']\n",
            "pred:  1 label:  1 ['data', 'on', 'current', 'hospitalizations', 'has', 'been', 'unstable', 'since', 'july', '15', '.', 'written', 'up', 'everything', 'we', 'know', 'about', 'problems', 'hospitals', 'and', 'states', 'are', 'having', 'and', 'about', 'some', 'unexpected', 'discrepancies', 'in', 'state', 'and', 'federal', '.']\n",
            "pred:  1 label:  1 ['daily', 'update', 'allows', 'us', 'to', 'track', 'individual', '24-hour', 'periods', '.', 'there', 'is', 'also', 'a', 'tab', 'that', 'provides', 'updated', 'numbers', 'throughout', 'day', '.', 'a', 'quick', 'reminder', 'about', 'our']\n",
            "pred:  0 label:  0 ['vaccine', 'has', 'been', 'developed', 'that', 'cures', 'in', 'just', 'three', 'hours', '.', '_']\n",
            "pred:  1 label:  1 ['travel', 'and', 'hospitality', 'stocks', 'have', 'been', 'hit', 'by', 'fears', 'of', 'a', 'second', 'lockdown', 'as', 'number', 'of', 'uk', 'coronavirus', 'cases', 'increase', '.']\n",
            "pred:  0 label:  0 ['two', 'cyclists', 'from', 'cordoba', 'skip', 'quarantine', '.']\n",
            "pred:  1 label:  1 ['cytokine', 'storm', 'of', 'immune', 'system', 'does', 'not', 'appear', 'to', 'play', 'a', 'major', 'role', 'in', 'more', 'severe', 'outcomes', 'according', 'to', 'some', 'unexpected', 'new', 'findings', '.', '.']\n",
            "pred:  0 label:  0 ['.', '.', 'for', 'efficient', 'spreading', 'in', 'human', 'population', 'compared', 'to', 'other', 'lineage', 'b', '.']\n",
            "pred:  1 label:  1 ['rt', '.', 'dr', '.', 'and', 'i', 'discuss', 'strategies', 'to', 'fight', 'vaccines', 'monoclonal', 'antibodies', 'convalescent']\n",
            "pred:  0 label:  0 ['stoner', 'surprised', 'to', 'learn', 'britain', 'is', 'on', 'lockdown', '.', '.']\n",
            "pred:  1 label:  1 ['new', 'report', 'in', 'found', 'higher', 'percentages', 'of', 'people', 'in', 'racial/ethnic', 'minority', 'groups', 'who', 'died', 'of', 'were', 'under', '65', '.', 'read', 'report', 'to', 'learn', '.', '.', '.']\n",
            "pred:  0 label:  0 ['bill', 'gates', 'helped', 'fund', 'a', 'patent', 'for', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['video', 'shows', 'a', 'doctor', 'from', 'breach', 'candy', 'hospital', 'of', 'mumbai', 'claiming', 'that', 'if', 'you', 'can', 'hold', 'your', 'breath', 'for', 'a', 'longer', 'period', 'without', 'discomfort', 'you', 'have', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['video', 'shows', 'dean', 'of', 'kasturba', 'hospital', 'telling', 'gargling', 'with', 'salt', 'water', 'kills', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['man', 'know', 'why', 'he', 'bothers', '.']\n",
            "pred:  0 label:  0 ['weinstein', 'to', 'be', 'streamed', '24/7', 'while', 'suffering', 'from', 'to', 'boost', 'morale', '.']\n",
            "pred:  1 label:  1 ['cabinet', 'will', 'again', 'review', 'settings', 'of', 'alert', 'level', '2', 'on', '8', 'june', 'and', 'have', 'agreed', 'that', 'no', 'later', 'than', '22', 'june', '4', 'weeks', 'from', 'today', 'they', 'will', 'consider', 'a', 'move', 'to', 'alert', 'level', '1', '.']\n",
            "pred:  0 label:  0 ['ivanka', 'trump', 'fears', 'that', 'her', 'daddy', 'may', 'have', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['an', 'image', 'has', 'been', 'shared', 'thousands', 'of', 'times', 'in', 'multiple', 'posts', 'on', 'twitter', 'and', 'facebook', 'which', 'claim', 'it', 'shows', 'overgrown', 'shrubs', 'at', 'a', 'theme', 'park', 'in', 'malaysia', 'during', 'a', 'coronavirus', 'lockdown', '.']\n",
            "pred:  1 label:  1 ['a', 'look', 'at', 'number', 'of', 'tests', 'per', '100000', 'people', 'in', 'state', '.', '.']\n",
            "pred:  0 label:  0 ['let', 'me', 'be', 'donald', 'trump', 'knew', 'about', 'threat', 'coronavirus', 'posed', 'and', 'failed', 'to', 'act', '.', 'one', 'of', 'most', 'unjustifiable', 'failures', 'of', 'presidential', 'leadership', 'in', 'american', 'history', '.', '.']\n",
            "pred:  1 label:  1 ['rt', '5', 'ways', 'to', 'help', 'cope', 'with', 'stress', 'during', 'pause', '.', 'breathe', '.', 'notice', 'how', 'you', 'feel', 'take', 'breaks', 'from']\n",
            "pred:  0 label:  0 ['holy', 'was', 'banned', 'in', 'china', 'now', 'ban', 'has', 'been', 'lifted', 'chinese', 'muslims', 'are', 'allowed', 'to', 'read', 'holy', 'quran']\n",
            "pred:  0 label:  0 ['pm', 'modi', 'said', 'that', '1', 'crore', 'corona', 'patients', 'have', 'been', 'treated', 'for', 'free']\n",
            "pred:  0 label:  0 ['phase', '3', 'means', 'that', 'if', 'mexico', 'reaches', 'phase', '4', 'all', 'hospitals', 'that', 'are', 'helping', 'older', 'people', 'are', 'going', 'to', 'let', 'them', 'die', 'to', 'give', 'care', 'to', 'young', '.']\n",
            "pred:  1 label:  1 ['people', 'have', 'recovered', 'more', 'than', 'active', 'cases', '.', 'with', 'this', 'recovery', 'rate', 'amongst', 'patients', 'has', 'crossed', '.', 'today', '.', 'details', ':', '.']\n",
            "pred:  0 label:  0 ['federal', 'bureau', 'of', 'investigation', 'arrested', 'billionaire', 'bill', 'gates', 'on', 'charges', 'of', 'biological', 'terrorism', 'after', 'he', 'allegedly', 'created', 'novel', 'coronavirus', 'virus', 'that', 'causes', '.']\n",
            "pred:  1 label:  1 ['boris', 'johnson', 'reveals', 'new', 'rules', 'on', 'face', 'masks', 'fines', 'pubs', 'and', 'working', 'from', 'home', '.']\n",
            "pred:  0 label:  0 ['uk', 'prepares', 'for', 'full', 'lockdown', 'after', 'hundreds', 'of', 'new', 'cases', '—', 'wuhan', 'super', 'virus', 'updates', '.', '”']\n",
            "pred:  1 label:  1 ['upward', 'trend', 'holds', 'for', 'death', 'data', '.', 'though', 'data', 'is', 'less', 'smooth', 'between', 'june', '10', 'and', 'july', '10', 'seven-day', 'average', 'of', 'deaths', 'reported', 'by', 'arizona', 'california', 'florida', 'south', 'carolina', 'and', 'texas', 'has', 'almost', 'doubled', '.', '.']\n",
            "pred:  0 label:  0 ['weave', 'and', 'lace', 'fronts', 'manufactured', 'in', 'china', 'may', 'contain', 'coronavirus', '.', '”']\n",
            "pred:  0 label:  0 ['lot', 'of', 'young', 'bankers', 'are', 'affected', 'by', 'corona', 'virus', 'and', 'soon', 'banks', 'can', 'become', 'biggest', 'outbreak', 'places', '.', 'these', 'young', 'ppl', 'can', 'become', 'threat', 'to', 'their', 'elders', 'at', 'home', '.', 'its', 'high', 'time', 'that', 'should', 'think', 'of', 'saving', 'these', 'lives', 'please', '.', '.']\n",
            "pred:  0 label:  0 ['michigan', 'governor', 'reopened', 'traverse', 'city', '?', '?', '?', 'so', 'she', 'could', 'have', 'her', 'open', 'house', '.', '?', '?']\n",
            "pred:  0 label:  0 ['dr', '.', 'li', 'wenliang', 'discovered', 'coffee', 'can', 'cure', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['one', 'other', 'u', '.', 's', '.', 'positive', 'rate', 'has', 'remained', 'around', 'throughout', 'april', '.', 'positive', 'rate', 'for', 'reported', 'tests', 'dropped', 'to', '.', 'one', 'day', 'make', 'a', 'trend', 'but', 'good', 'to', 'see', 'a', 'lower', 'floor', '.']\n",
            "pred:  1 label:  1 ['together', 'we', 'have', 'kept', 'out', 'for', '102', 'days', '-', 'longer', 'than', 'any', 'other', 'country', '.', 'we', 'can', 'do', 'all', 'of', 'that', 'again', '.', 'if', 'in', 'auckland', 'please', 'stay', 'at', 'home', 'with', 'your', 'bubble', '.', 'take', 'care', '.', 'got', 'this', 'new']\n",
            "pred:  0 label:  0 ['claim', 'that', 'rep', '.', 'alexandria', 'tweeted', 'that', 'u', '.', 's', '.', 'governors', 'should', 'extend', 'restrictions', 'on', 'businesses', 'after', 'november', 'elections', '.', '”']\n",
            "pred:  1 label:  1 ['south', 'set', 'a', 'new', 'record', 'for', 'deaths', 'across', 'region', 'yesterday', 'at', '592', '.', 'today', 'south', 'reported', '725', 'deaths', '.', '.']\n",
            "pred:  1 label:  1 ['as', 'at', '11:25', 'pm', '21st', 'april', 'breakdown', 'of', 'cases', 'by', 'akwa']\n",
            "pred:  0 label:  0 ['channel', 'jon', 'snow', 'is', 'infected', 'with', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['how', 'much', 'is', 'he', 'stealing', 'from', 'us', 'besides', 'what', 'is', 'listed', '?', 'how', 'much', 'did', 'he', 'invest', 'in', 'hydro', 'chloroquine', 'before', 'touting', 'it', 'to', 'fools', 'as', 'a', 'cure', 'for', '?', 'how', 'much', 'did', 'he', 'steal', 'taking', 'from', 'states', 'and', 'reselling', 'them', '?', '1/2']\n",
            "pred:  0 label:  0 ['saddam', 'hussein', 'at', 'a', 'meeting', 'in', 'nineties', 'reported', 'that', 'us', 'threatened', 'to', 'quit', 'corona', 'virus', 'on', 'iraq']\n",
            "pred:  0 label:  0 ['lick', 'a', 'zebra', 'for', 'more', 'than', '3', '–', 'government', 'coronavirus', 'advice', 'enters', 'surreal', 'stage', '.', '.']\n",
            "pred:  1 label:  1 ['and', 'launch', 'law', 'lab', 'a', 'platform', 'that', 'shares', 'legal', 'documents', 'from', 'over', '190', 'countries', 'across', 'to', 'support', 'global', 'response', '.', 'more', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['studies', 'show', 'that', 'people', 'who', 'get', 'positive', 'tests', 'quite', 'late', 'into', 'illness', 'are', 'not', 'infectious', '.', 'if', 'they', 'test', 'positive', 'on', 'day', '12', 'visit', 'are', 'prolonged', 'until', 'we', 'are', 'sure', 'they', 'have', 'recovered', '.']\n",
            "pred:  1 label:  1 ['of', '80', 'cases', 'linked', 'to', 'community', 'outbreak', '78', 'are', 'linked', 'to', 'auckland', 'cluster', 'and', '2', 'remain', 'under', 'maintenance', 'worker', 'at', 'hotel', 'facility', 'and', 'a', 'case', 'announced', 'yesterday', 'which', 'has', 'been', 'reclassified', 'as', 'under', 'investigation', '.']\n",
            "pred:  0 label:  0 ['cases', 'are', '?', '?', '?', 'up', 'only', 'because', 'of', 'our', 'big', 'number', 'testing', '.', '?', '?']\n",
            "pred:  1 label:  1 ['kia', 'ora', 'liz', 'it', 'is', 'up', 'to', 'each', 'cafe', 'to', 'decide', 'this', '.', 'we', 'can', 'imagine', 'you', 'will', 'be', 'allowed', 'to', 'use', 'them', 'again']\n",
            "pred:  1 label:  1 ['data', 'indicate', 'novel', 'coronavirus', 'is', 'capable', 'of', 'spreading', 'easily', 'and', 'sustainably', 'from', 'person', 'to', 'person', '.', 'because', 'it', 'is', 'a', 'new', 'virus', 'there', 'is', 'essentially', 'no', 'immunity', 'against', 'in', 'population', '.', '.']\n",
            "pred:  0 label:  0 ['grocery', ',', 'milk', ',', 'chemist', ',', 'vegetable', 'will', 'be', 'open', 'from', 'tomorrow', '-', '24th', 'march', '2020']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '4', 'million', 'tests', 'up', 'from', 'yesterday', 'a', 'bit', 'lower', 'than', 'april', 'average', 'of', 'tests', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'and', 'not', 'all', 'states', 'report', 'all', 'tests', '.', 'for', 'details', '.']\n",
            "pred:  1 label:  1 ['data', 'today', 'is', 'confusing', '.', 'after', 'days', 'of', 'huge', 'test', 'volumes', 'back', 'to', 'levels', 'of', 'earlier', 'plateau', '.', 'fewest', 'new', 'cases', 'were', 'confirmed', 'of', 'any', 'day', 'in', 'april', 'but', 'positive', 'rate', 'was', 'higher', 'than', 'in', 'last', 'few', 'days', 'too', '.']\n",
            "pred:  0 label:  0 ['real', 'amount', 'of', 'fatalities', 'to', '200', ',', '000', 'american', 'lives', '.', 'not', '10', ',', '000', 'you', 'and', 'doctor', 'who', 'took', 'anti-malarial', 'drugs', 'to', 'show', 'solidarity', 'with', 'have', 'stated', '.', 'lies', 'cost']\n",
            "pred:  1 label:  1 ['rt', 'bulletin', 'india', 'reaches', 'another', 'record', 'of', 'highest', 'single', 'day', 'recoveries', 'recover', 'in', 'last', '24', 'hours']\n",
            "pred:  0 label:  0 ['a', 'new', 'vaccine', 'to', 'cure', 'coronavirus', 'has', 'been', 'developed', 'by', 'u', '.', 's', '.', 'scientists', '.']\n",
            "pred:  1 label:  1 ['yeah', '.', '.', '.', 'i', 'will', 'believe', 'this', 'when', 'i', 'see', 'it', '.', 'no', 'evidence', 'has', 'that', 'capacity', '.', 'especially', 'when', 'of', 'tests', 'are', 'coming', 'back', 'positive', '.']\n",
            "pred:  0 label:  0 ['a', 'youtube', 'video', 'shared', 'thousands', 'of', 'times', 'claims', 'that', 'two', 'children', 'died', 'from', 'a', 'novel', 'coronavirus', 'vaccine', 'in', 'guinea', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'bill', 'gates', 'explains', 'that', 'vaccine', 'will', 'use', 'experimental', 'technology', 'and', 'permanently', 'alter', 'your', 'dna', '?', '?']\n",
            "pred:  1 label:  1 ['looking', 'at', 'case', 'data', 'there', 'is', 'now', 'a', 'fairly', 'clear', 'low', 'plateau', 'where', '7-day', 'average', 'stayed', 'between', 'cases', 'per', 'day', '.', 'this', 'extended', 'from', 'to', 'so', 'from', '29', 'ago', 'to', '7', 'days', 'ago', '.', '.']\n",
            "pred:  0 label:  0 ['social', 'services', 'to', 'investigate', 'all', 'parents', 'with', 'coronavirus', 'symptoms', 'who', 'travel', '250', 'miles', 'to', 'get', 'help', '.', '.']\n",
            "pred:  0 label:  0 ['seventy-three', 'japanese', 'police', 'officers', 'have', 'been', 'infected', 'with', 'after', 'handling', 'infected', 'bodies', '.']\n",
            "pred:  1 label:  1 ['543', 'new', 'cases', 'of', ';', 'confirmed', 'discharged', '813', 'deaths', '.']\n",
            "pred:  0 label:  0 ['hospitals', 'are', 'empty', 'and', 'it', 'proves', 'that', 'epidemic', 'exist', '.']\n",
            "pred:  1 label:  1 ['rt', '.', 'is', 'creating', 'a', 'legal', 'duty', 'to', 'if', 'someone', 'tests', 'positive', 'for', '.', 'is', 'this', 'a', 'sensible', 'policy', '?', \"'\"]\n",
            "pred:  0 label:  0 ['three', 'gorges', 'dam', 'faces', 'severe', 'flooding', 'as', 'yangtze', 'overflows', ',', 'yangtze', 'river', 'flood', 'in', 'streets', '3', 'bridges', 'in', 'this', 'flood', '.']\n",
            "pred:  1 label:  1 ['three', 'stories', 'to', 'read', 'this', 'lockdown', 'restrictions', 'begin', 'in', 'north', 'east', '.', 'national', 'lockdown', 'in', 'october', 'proposed', 'by', 'top', 'scientists', '.', 'in', 'denial', 'about', '.', '.']\n",
            "pred:  1 label:  1 ['mandatory', '14', 'days', 'in', 'isolation', 'is', 'key', 'part', 'of', 'our', 'border', 'control', '.', 'routine', 'testing', 'that', 'we', 'added', 'last', 'week', 'is', 'an', 'additional', 'measure', '.']\n",
            "pred:  0 label:  0 ['an', 'image', 'of', 'a', 'doctor', 'who', 'found', 'cure', 'for', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['now', 'passed', '100', 'days', 'without', 'community', 'transmission', 'but', 'testing', 'remains', 'one', 'of', 'best', 'ways', 'to', 'ensure', 'no', 'undetected', 'community', 'transmission', 'in', 'new', 'zealand', '.', 'we', 'need', 'everyone', 'to', 'play', 'their', 'part', 'in', 'that', '.']\n",
            "pred:  1 label:  1 ['stepping', 'back', 'infection', 'and', 'outcomes', 'data', 'and', 'are', 'not', 'moving', 'together', '.', 'deaths', 'and', 'hospitalizations', 'continue', 'to', 'fall', ';', 'cases', 'fell', 'but', 'now', 'seem', 'to', 'be', 'ticking', 'up', '.']\n",
            "pred:  1 label:  1 ['there', 'are', 'currently', 'people', 'in', 'managed', 'isolation', 'and', 'quarantine', '.', 'our', 'current', 'effective', 'capacity', 'is', '.', 'this', 'gives', 'us', 'an', 'excess', 'capacity', 'of', '.', 'over', 'next', 'week', 'we', 'are', 'projecting', 'arrivals', 'and', '1164', 'departures', 'from', 'our', 'facilities', '.']\n",
            "pred:  0 label:  0 ['to', 'every', 'military', 'family', 'in', 'our', 'thank', 'you', ',', 'thank', 'you', ',', 'thank', 'you', '.', '.']\n",
            "pred:  1 label:  1 ['10', 'account', 'for', 'nearly', 'of', 'new', 'recovered', 'cases', '.', 'maharashtra', 'has', 'maintained', 'this', 'lead', 'with', 'new', 'recoveries', '.', '.']\n",
            "pred:  0 label:  0 ['if', 'you', 'donate', 'blood', 'have', 'to', 'test', 'for', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['hundreds', 'of', 'doctors', 'start', 'their', 'residencies', 'on', 'time', 'leaving', 'front', 'line', 'understaffed', '.', '.']\n",
            "pred:  0 label:  0 ['although', 'germany', 'closes', 'its', 'borders', 'refugees', 'and', 'migrants', 'may', 'still', 'come', 'in', 'and', 'they', 'will', 'not', 'be', 'medically', 'checked', 'for', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['industrial', 'bleach', 'being', 'sold', 'on', 'amazon', 'as', 'mineral', 'of', '.', 'please', 'take', 'time', 'to', 'read', 'our', 'story', 'and', 'if', 'you', 'can', ',', 'please', 'if', 'not', 'possible', 'for', 'you', ',', 'please', 'share', 'our', '.']\n",
            "pred:  1 label:  1 ['we', 'are', 'proud', 'to', 'announce', 'that', 'india', 'has', 'realised', 'pm', 'vision', 'of', 'conducting', '10', 'lakh', 'tests', 'per', 'this', 'is', 'a', 'significant', 'milestone', 'in', 'our', 'fight', 'against', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'oklahoma', 'coronavirus', 'test', 'rate', 'after', 'trump', 'tulsa', 'rally', '.', '?', '?']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '13', '.', '8', 'million', 'tests', 'up', 'from', 'yesterday', '.', '5', 'states', 'showed', 'a', 'decline', 'in', 'total', 'tests', 'due', 'to', 'separating', 'out', 'their', 'antibody', 'mi', 'mo', 'ms', 'tx', 'wv', '.', 'details', 'in', 'this', 'thread', '.', '.']\n",
            "pred:  1 label:  1 ['biggest', 'risk', 'to', 'us', 'is', 'people', 'coming', 'into', 'country', 'from', 'overseas', 'which', 'is', 'why', 'those', 'people', 'must', 'stay', 'in', 'managed', 'isolation', 'or', 'quarantine', 'for', 'at', 'least', '14', 'days', '.']\n",
            "pred:  1 label:  1 ['recovery', 'rate', 'improves', 'to', '78', '.', 'as', 'on', 'september', '18', '2020', 'improvement', 'in', 'recovery', 'rate', 'since', 'initiation', 'on', 'march', '25', '2020', 'via', '.']\n",
            "pred:  1 label:  1 ['t-cell', 'response', 'may', 'offer', 'a', 'more', 'precise', 'measure', 'of', 'potential', 'long-term', 'immune', 'protection', 'from', 'compared', 'with', 'antibody', 'assays', '.', '.']\n",
            "pred:  0 label:  0 ['treatment', 'cures', 'full', 'stop', '.', 'we', 'dont', 'need', 'a', 'if', 'they', 'push', 'an', 'untested', 'vaccine', 'early', ',', 'or', 'if', 'they', 'push', 'stay', 'at', 'home', 'until', 'we', 'have', 'a', 'vaccine', ',', 'then', 'this', 'is', 'not', 'about', 'virus', '.']\n",
            "pred:  0 label:  0 ['president', 'of', 'united', 'states', 'of', 'america', 'donald', 'trump', 'has', 'been', 'tested', 'positive', 'for', '.']\n",
            "pred:  0 label:  0 ['fit', 'people', 'are', 'immune', 'to', 'novel', 'coronavirus']\n",
            "pred:  1 label:  1 ['of', '2159', 'people', 'who', 'left', 'managed', 'isolation', 'facilities', 'during', 'this', 'period', '1284', 'people', 'have', 'now', 'been', 'contacted', 'and', 'have', 'tested', 'negative', 'for', ';', '800', 'of', 'those', 'were', 'tested', 'before', 'leaving', 'managed', 'isolation', 'and', 'remaining', '484', 'were', 'tested', 'after', 'departure', 'from', 'facility', '.']\n",
            "pred:  1 label:  1 ['we', 'are', 'shipping', '.', '7', 'million', 'dollars', 'worth', 'of', 'personal', 'protective', 'equipment', 'items', 'to', 'support', 'and', 'humanitarian', 'supplies', 'that', 'were', 'destroyed', 'by']\n",
            "pred:  0 label:  0 ['pib', 'has', 'clarified', 'that', 'cover', 'related', 'deaths', ',', 'while', 'covers', 'deaths', 'with', 'certain', 'conditions', '.']\n",
            "pred:  1 label:  1 ['police', 'commissioner', 'andrew', '200', '-', '350', 'reports', 'of', 'breaches', 'per', 'day', 'than', '700', '-', '1000', 'reports', 'under', 'alert', 'level', '983', 'breaches', 'were', 'reported', 'to', 'police']\n",
            "pred:  1 label:  1 ['our', 'laboratories', 'processed', 'tests', 'our', 'highest', 'day', 'of', 'testing', 'since', 'we', 'began', '.', 'this', 'includes', 'testing', 'in', 'managed', 'isolation', 'facilities', 'as', 'well', 'as', 'testing', 'across', 'community', '.']\n",
            "pred:  0 label:  0 ['president', 'trump', 'said', 'of', 'cases', 'are', 'totally', 'harmless', '.', 'false', '.', 'from', 'deaths', 'to', 'hospitalizations', 'to', 'a', 'growing', 'list', 'of', 'ailments', 'that', 'hit', 'people', 'with', 'even', 'mild', 'cases', 'numbers', 'say', 'his', 'claim', 'is', 'wrong', '.', '.', '.']\n",
            "pred:  1 label:  1 ['issues', 'new', 'guidelines', 'opens', 'up', 'more', 'activities', 'outside', 'containment', 'zones', 'enforcement', 'of', 'in', 'containment', 'zones', 'till', '30th', 'september', '2020', '.', '.']\n",
            "pred:  1 label:  1 ['latest', 'cdc', 'report', 'shows', 'that', 'percentage', 'of', 'people', 'testing', 'positive', 'for', 'and', 'percent', 'of', 'medical', 'visits', 'for', 'symptoms', 'similar', 'to', 'are', 'decreasing', 'nationally', '.', 'learn', '.', '.', '.']\n",
            "pred:  0 label:  0 ['messages', 'that', 'claim', 'that', 'government', 'knows', 'where', 'we', 'are', 'at', 'all', 'times', 'thanks', 'to', 'an', 'application', 'that', 'google', 'has', 'installed', 'on', 'our', 'phones', '.']\n",
            "pred:  0 label:  0 ['claim', 'that', 'fox', 'news', 'said', 'that', 'trump', 'tested', 'positive', 'for', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'corona', 'virus', 'florida', 'man', 'arrested', 'for', 'robbery', 'using', 'cough', 'as', 'a', 'weapon', '.', '?', '?']\n",
            "pred:  1 label:  1 ['rt', 'india', 'scales', 'another', 'peak', '.', 'in', 'last', '24', 'hours', '12', 'lakh', 'tests', 'were', 'conducted', 'across', 'country', 'which']\n",
            "pred:  1 label:  1 ['we', 'think', 'better', 'data', 'is', 'better', 'for', 'everyone', '.', 'it', 'informs', 'where', 'governments', 'spend', 'money', 'allocate', 'resources', 'and', 'manage', 'a', 'crisis', '.', 'residents', 'can', 'make', 'better', 'decisions', '.', 'data', 'is', 'not', 'a', 'panacea', 'but', 'it', 'can', 'help', 'nebraska', 'and', 'others', 'plan', 'for', 'rest', 'of', 'this', 'future', 'ones', '.']\n",
            "pred:  1 label:  1 ['not', 'necessarily', 'subsided', 'but', 'flattened', 'in', 'wa', 'ca', 'and', '.', 'waves', 'are', 'overlapping', 'though', '.']\n",
            "pred:  1 label:  1 ['here', 'are', 'states', 'that', 'reported', 'over', '500', 'cases', 'today', '.', 'mississippi', 'had', 'data', 'reporting', 'problems', 'for', 'days', 'so', 'their', 'number', 'represents', 'data', 'from', '.', '.']\n",
            "pred:  0 label:  0 ['corona', 'patients', 'have', 'started', 'appearing', 'in', 'public', 'places', 'on', 'streets', 'of', 'pune', 'situation', 'is', 'dire', ',', 'kindly', 'please', 'take', 'care', 'of', 'yourself', 'and', 'your', 'family', 'stay', 'safe', '.', '.', '.', '.']\n",
            "pred:  0 label:  0 ['u', '.', 's', '.', 'senator', 'lindsey', 'graham', 'says', 'president', 'trump', 'knows', 'more', 'about', 'medicine', 'and', 'specifically', 'viruses', 'than', 'any', 'of', 'so-called', 'scientists', 'at', 'nih', 'and', 'cdc', '.', '”']\n",
            "pred:  1 label:  1 ['all', 'others', 'had', 'completed', '14', 'days', 'in', 'managed', 'isolation', 'and', '1326', 'people', 'have', 'now', 'been', 'contacted', 'and', 'have', 'tested', 'negative', 'for', 'which', 'is', 'over', 'of', 'this', 'cohort', '.', 'a', 'further', '395', 'people', 'have', 'been', 'spoken', 'with', 'and', 'referred', 'for', 'testing', '.']\n",
            "pred:  0 label:  0 ['police', 'patrols', 'are', 'everywhere', 'with', 'steel', 'forks', ',', 'iron', 'bars', 'and', 'guns', '.', '.', '.', 'is', 'this', 'way', 'to', 'stopped', 'outbreak', '?', 'no', 'way', '.', '.', '.', '!', '.']\n",
            "pred:  1 label:  1 ['signs', 'and', 'symptoms', 'of', '?', 'if', 'you', 'or', 'your', 'loved', 'one', 'has', 'an', 'infection', 'not', 'getting', 'better', 'or', 'is', 'getting', 'worse', 'act', 'fast', 'and', 'get', 'medical', 'care', 'immediately', '.', '.', '.']\n",
            "pred:  0 label:  0 ['prime', 'minister', 'says', 'that', 'curve', 'is', 'beginning', 'to', '.', 'but', 'is', 'it', 'and', 'how', 'do', 'we', 'know', '?', 'rmit', 'abc', 'fact', 'check', 'explores', 'coronavirus', '.', '.']\n",
            "pred:  0 label:  0 ['no', ',', 'a', 'reflection', 'of', 'this', 'bullshit', 'that', 'yes', ',', 'there', 'is', 'a', 'virus', 'but', 'no', ',', 'not', 'as', 'deadly', 'as', 'people', 'think', 'and', 'being', 'blown', 'out', 'of', 'proportion', 'to', 'shut', 'everything', 'down', 'to', 'blame', 'public', 'for', 'their', 'own', 'demise', '.', '.', '.', 'age', 'old', 'tory', 'trick', '.', '.', '.']\n",
            "pred:  1 label:  1 ['two', 'notes', 'in', 'numbers', 'washington', 'state', 'revealed', 'they', 'had', 'been', 'counting', 'antibody', 'tests', 'in', 'its', 'numbers', 'so', 'their', 'total', 'number', 'of', 'tests', 'will', 'drop', '.', 'yesterday', 'we', 'accidentally', 'counted', 'antibody', 'tests', 'in', 'its', 'total', '.', 'correcting', 'data', 'today', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'states', 'reported', 'tests', 'and', 'cases', '.', '1088', 'deaths', 'were', 'reported', 'today', 'about', '100', 'down', 'from', 'last', 'saturday', '.', '.']\n",
            "pred:  0 label:  0 ['says', 'bbc', 'prematurely', 'reported', '?', '?', '?', 'ghislaine', 'maxwell', 'moved', 'to', 'intensive', 'care', 'as', 'coronavirus', 'symptoms', 'worsen', '.', '?', '?']\n",
            "pred:  0 label:  0 ['doctors', 'warn', 'loss', 'of', 'sense', 'of', 'taste', 'caused', 'by', 'coronavirus', 'can', 'result', 'in', 'people', 'buying', 'nickelback', 'albums', '.', '.']\n",
            "pred:  0 label:  0 ['video', 'claims', 'that', '19', 'is', 'a', 'bacteria', 'and', 'can', 'be', 'treated', 'by', 'antibiotics']\n",
            "pred:  0 label:  0 ['media', 'why', 'newspapers', 'reporting', 'about', 'virus', '.']\n",
            "pred:  0 label:  0 ['a', 'facebook', 'post', 'claiming', 'that', 'million', 'who', 'died', 'during', 'second', 'wave', 'of', 'spanish', 'flu', 'had', 'just', 'got', 'a', 'vaccine', '.', 'post', 'compares', 'this', 'situation', 'with', 'new', 'coronavirus', 'and', 'warns', 'people', 'against', 'getting', 'a', 'vaccine', 'for', 'when', 'it', 'will', 'be', 'ready', '.']\n",
            "pred:  1 label:  1 ['rt', 'handwashing', 'is', 'one', 'of', 'best', 'ways', 'to', 'protect', 'yourself', ';', 'your', 'family', 'from', 'getting', 'sick', 'with', '.', 'learn', 'when', ';', 'how']\n",
            "pred:  0 label:  0 ['drinking', 'alcohol', 'kills', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['it', 'have', 'to', 'be', 'this', 'bad', '.', 'donald', 'trump', 'utterly', 'failed', 'to', 'prepare', 'for', 'this', 'pandemic', 'and', 'delayed', 'in', 'taking', 'necessary', 'steps', 'to', 'safeguard', 'our', 'nation', 'against', 'scenario', 'we', 'are', 'now', 'living', '.', '.']\n",
            "pred:  0 label:  0 ['says', 'n', '.', 'y', '.', 'gov', '.', 'andrew', 'cuomo', '?', '?', '?', 'rejected', 'buying', 'recommended', '16', ',', '000', 'ventilators', 'in', '2015', 'for', 'pandemic', ',', 'for', 'a', 'pandemic', ',', 'established', 'death', 'panels', 'and', 'lotteries', 'instead', '.', 'so', ',', 'he', 'had', 'a', 'chance', 'to', 'buy', ',', 'in', '2015', ',', '16', ',', '000', 'ventilators', 'at', 'a', 'very', 'low', 'price', ',', 'and', 'he', 'turned', 'it', 'down', '.', '?', '?']\n",
            "pred:  0 label:  0 ['vaccine', 'against', 'new', 'coronavirus', 'has', 'been', 'developed', 'in', 'ukraine', '.']\n",
            "pred:  1 label:  1 ['of', '2159', 'people', 'who', 'left', 'managed', 'isolation', 'facilities', 'between', 'june', '9-16', '1228', 'people', 'have', 'been', 'contacted', 'and', 'have', 'tested', 'negative', 'for', ';', '800', 'of', 'those', 'were', 'tested', 'before', 'leaving', 'managed', 'isolation', 'and', 'remaining', '428', 'were', 'tested', 'after', 'departure', 'from', 'facility', '.']\n",
            "pred:  0 label:  0 ['novel', 'coronavirus', 'was', '?', '?', '?', 'man-made', '.', '?', '?']\n",
            "pred:  1 label:  1 ['as', 'at', '09:30', 'pm', '5th', 'april', 'there', 'are', '232', 'confirmed', 'cases', '33', 'discharged', '5', 'deaths', 'for', 'a', 'breakdown', 'of', 'cases', 'by', '.', '120', '47', '20', '9', '9', '6', 'akwa', '5', '5', '4', '2', '2', '1', '1']\n",
            "pred:  0 label:  0 ['words', 'of', 'a', 'president', 'matter', '—', 'and', 'donald', 'trump', 'has', 'used', 'his', 'to', 'downplay', ',', 'pass', 'blame', 'onto', 'others', ',', 'and', 'mislead', 'american', 'people', '.', 'i', 'will', 'do', 'opposite', '.', '.']\n",
            "pred:  1 label:  1 ['a', 'number', 'of', 'vaccines', 'are', 'now', 'in', 'final', 'stage', 'of', 'clinical', 'trials', 'and', 'we', 'all', 'hope', 'have', 'multiple', 'successful', 'candidates', 'that', 'are', 'both', 'safe', 'and', 'effective', '.']\n",
            "pred:  0 label:  0 ['a', 'post', 'claims', 'compulsory', 'violates', 'principles', 'of', 'bioethics', 'that', 'coronavirus', 'exist', 'that', 'pcr', 'test', 'returns', 'many', 'false', 'positives', 'and', 'that', 'influenza', 'vaccine', 'is', 'related', 'to', '.']\n",
            "pred:  0 label:  0 ['a', 'vaccine', 'that', 'is', 'being', 'developed', 'in', 'china', 'and', 'will', 'be', 'tested', 'in', 'brazil', 'is', 'being', 'funded', 'by', 'bill', 'gates', '.']\n",
            "pred:  0 label:  0 ['multiple', 'posts', 'shared', 'repeatedly', 'on', 'facebook', 'and', 'twitter', 'claim', 'that', 'a', 'hong', 'kong', 'medical', 'lab', 'has', 'warned', 'novel', 'coronavirus', 'can', 'remain', 'viable', 'on', 'fruits', 'and', 'vegetables', 'for', '12', 'hours', 'therefore', 'people', 'should', 'over', 'fears', 'of', 'contracting', '.']\n",
            "pred:  0 label:  0 ['400', 'are', 'stuck', 'at', 'vaishno', 'devi', 'due', 'to', '.']\n",
            "pred:  1 label:  1 ['number', 'of', 'cases', 'is', 'rising', 'in', 'all', 'uk', 'regions', 'but', 'compared', 'to', 'other', 'areas', 'capital', 'appears', 'to', 'be', 'controlling', 'virus', 'better', 'than', 'some', 'regions', 'that', 'had', 'similar', 'or', 'higher', 'infection', 'rates', 'in', 'april', 'and', 'may', '.']\n",
            "pred:  1 label:  1 ['a', 'second', 'lockdown', 'would', 'be', 'a', 'failure', '-', 'not', 'an', 'act', 'of', 'sir', 'keir', 'starmer', 'has', 'said', '.']\n",
            "pred:  0 label:  0 ['minister', 'of', 'state', 'for', 'ayush', 'shripad', 'naik', 'held', 'a', 'press', 'conference', 'claiming', 'that', 'prince', 'charles', 'followed', 'rules', 'of', 'ayurveda', 'which', 'is', 'why', 'he', 'is', 'recovering', 'so', 'soon', '.']\n",
            "pred:  1 label:  1 ['.', 'mortality', 'risk', 'in', 'bystander', 'cpr', 'events', '.', '.']\n",
            "pred:  0 label:  0 ['or', 'nyc', 'millions', 'will', 'be', 'infected', 'by', 'and', 'up', 'to', '200', ',', '000', 'will', 'die', 'in', 'u', '.', 's', '.', ',', 'top', 'expert', 'anthony', 'fauci', 'warned', 'sunday', ',', 'bloomberg', 'reports', '.', '.']\n",
            "pred:  0 label:  0 ['new', 'mask', 'exemption', 'card', 'indicates', 'medical', 'condition', 'of', 'being', 'an', 'asshole', '.', '.']\n",
            "pred:  0 label:  0 ['there', 'are', '59', 'positive', 'coronavirus', 'cases', 'in', 'nagpur', 'along', 'with', 'three', 'doctors', ',', 'one', 'of', 'whom', 'is', 'on', 'ventilator']\n",
            "pred:  1 label:  1 ['these', 'people', 'have', 'been', 'assessed', 'as', 'presenting', 'a', 'very', 'low', 'risk', 'due', 'to', 'nature', 'of', 'their', 'exemption', 'adherence', 'to', 'their', 'required', 'protocols', 'and', 'negative', 'test', 'results', 'of', 'people', 'associated', 'with', 'their', 'bubble', '.']\n",
            "pred:  1 label:  1 ['auckland', 'funerals', 'or', 'can', 'host', 'up', 'to', '100', 'people', 'and', 'no', 'longer', 'need', 'to', 'be', 'registered', 'with', '.', 'from', 'midnight', 'tonight', 'rest', 'of', 'new', 'zealand', 'will', 'return', 'to', 'alert', 'level', '1', '.', 'alert', 'level', '1', 'is', 'not', 'alert', 'level', '0', '.']\n",
            "pred:  0 label:  0 ['countries', 'were', 'exporting', 'test', 'in', '2018']\n",
            "pred:  1 label:  1 ['some', 'veterinary', 'labs', 'help', 'with', 'human', 'testing', 'but', 'meet', 'demand', 'for', 'daily', 'tests', '.', '.']\n",
            "pred:  0 label:  0 ['only', '56', 'coronavirus', 'deaths', 'a', 'day', 'was', 'enough', 'to', 'collapse', 'world', 'economy', ',', 'even', 'though', 'thousands', 'of', 'people', 'die', 'daily', 'from', 'tuberculosis', ',', 'hepatitis', 'b', ',', 'and', 'other', 'diseases', '.']\n",
            "pred:  0 label:  0 ['china', 'concerned', 'western', 'authoritarian', 'government', 'will', 'cover', 'up', 'coronavirus', 'outbreaks']\n",
            "pred:  0 label:  0 ['china', 'pm', 'who', 'told', 'translate', 'quran', 'as', 'our', 'own', 'realised', 'that', 'only', 'one', 'way', 'to', 'protect', 'from', 'korona', 'virus', 'is', 'doing', 'to', 'allah', 'and', 'went', 'to', 'mosque', 'to', 'pray', ',', 'masha']\n",
            "pred:  0 label:  0 ['some', 'about', 'spraying', 'of', 'alcohol', 'or', 'chlorine', 'over', 'body', 'will', 'not', 'kill', 'viruses', ';', 'drinking', 'alcohol', 'smoking', 'or', 'eating', 'garlic', 'does', 'not', 'protect', 'you', 'from', ';', 'sesame', 'oil', 'does', 'not', 'kill', 'it', '.', 'a', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'was', 'called', 'a', 'by', 'who', ';', '2', '.', '9', 'million', 'were', 'infected', 'and', 'died', '.']\n",
            "pred:  1 label:  1 ['rt', 'take', 'personal', 'responsibility', 'to', 'help', 'slow', 'spread', 'of', 'so', 'that', 'you', 'can', 'say', '-', 'social']\n",
            "pred:  1 label:  1 ['there', 'are', 'currently', '5', 'people', 'receiving', 'hospital', 'level', 'care', '.', 'two', 'are', 'in', 'auckland', 'city', 'hospital', 'and', 'three', 'are', 'in', 'middlemore', 'hospital', '.', 'our', 'total', 'number', 'of', 'confirmed', 'cases', 'is', '1280', '.']\n",
            "pred:  1 label:  1 ['oxycontin', 'will', 'not', 'solve', 'opioid', 'epidemic', '.', '.']\n",
            "pred:  0 label:  0 [':', 'has', 'already', 'discovered', 'a', 'vaccine', '?', '.']\n",
            "pred:  1 label:  1 ['corresponding', 'figures', 'for', 'andhra', 'pradesh', 'and', 'karnataka', 'stand', 'at', 'and', '.', 'tamil', 'nadu', 'follows', 'with', 'while', 'uttar', 'pradesh', 'posted', 'recovery', 'of', 'patients', '.', 'details', ':', '.', '.']\n",
            "pred:  1 label:  1 ['hospitalized', 'patients', 'should', 'all', 'undergo', 'mri', '.', '.']\n",
            "pred:  1 label:  1 ['california', 'florida', 'and', 'texas', 'combined', 'accounted', 'for', 'of', 'all', 'new', 'cases', 'today', '.', '.']\n",
            "pred:  1 label:  1 ['though', 'positive', 'rate', 'has', 'been', 'declining', 'as', 'more', 'testing', 'capacity', 'comes', 'online', 'number', 'of', 'cases', 'being', 'confirmed', 'in', 'u', '.', 's', '.', 'each', 'day', 'continues', 'to', 'bounce', 'in', 'a', 'band', 'around', '30k', 'cases', '.', '.']\n",
            "pred:  1 label:  1 ['scaling', 'up', 'testing', 'is', 'key', 'in', 'response', 'enabling', 'timely', 'isolation', ';', 'treatment', 'of', 'cases', '.', 'currently', 'is', 'supporting', 'in', 'ongoing', 'community', 'testing', 'across', 'all', '27', 'lgas', 'in', 'state', '.', '.']\n",
            "pred:  1 label:  1 ['today', 'there', 'are', 'three', 'people', 'in', 'hospital', 'with', '–', 'one', 'is', 'in', 'isolation', 'on', 'a', 'ward', 'in', 'auckland', 'city', 'hospital', '.', 'two', 'are', 'in', 'icu', 'at', 'north', 'shore', 'and', 'waikato', 'hospitals', '.', 'we', 'are', 'sending', 'them', 'all', 'our', 'support', 'and', 'aroha', '.']\n",
            "pred:  0 label:  0 ['7', 'indian', 'states', 'saw', 'a', 'decrease', 'in', 'active', 'cases', 'in', 'last', '24', 'hours', '.', 'list', 'includes', 'big', 'names', 'like', ',', '.', '.']\n",
            "pred:  1 label:  1 ['rt', '.', 'has', 'decided', 'to', 'deploy', 'high', 'level', 'central', 'teams', 'to', 'four', 'states', 'of']\n",
            "pred:  1 label:  1 ['12', 'states', 'and', 'uts', 'have', 'registered', 'recovery', 'rate', 'more', 'than', 'national', 'average', '.', 'of', 'new', 'recovered', 'cases', 'are', 'from', '10', '.', 'details', ':', '.', '.']\n",
            "pred:  0 label:  0 ['videos', 'show', 'pigs', 'being', 'buried', 'in', 'china', 'due', 'to', 'coronavirus', 'outbreak', '.']\n",
            "pred:  1 label:  1 ['following', 'national', 'lead', '13', 'are', 'also', 'reporting', 'a', 'higher', 'number', 'of', 'new', 'recoveries', 'than', 'new', 'cases', '.', '.']\n",
            "pred:  0 label:  0 ['barely', 'kills', 'anyone', ',', '”', 'says', 'man', 'who', 'would', 'probably', 'be', 'pretty', 'annoyed', 'if', 'we', 'killed', 'him', '.', '.']\n",
            "pred:  1 label:  1 ['doctors', 'send', 'a', 'crystal', 'clear', 'message', 'about', 'importance', 'of', 'masks', 'for', 'prevention', '.', '.', 'via']\n",
            "pred:  0 label:  0 ['on', 'a', 'cheap', 'bourbon', 'to', 'gargle', 'my', 'throat', '.', 'best', 'virus', 'killers', 'in', 'world', ',', 'pure', 'battery', 'acid', ':)']\n",
            "pred:  0 label:  0 ['prince', 'harry', 'ordered', 'home', 'by', 'prince', 'charles', 'due', 'to', 'coronavirus', '?', '.']\n",
            "pred:  0 label:  0 ['andrés', 'fabián', 'hurtado', 'mayor', 'of', 'ibagué', 'says', 'his', 'city', 'lost', 'four', 'million', 'jobs', 'due', 'to', 'coronavirus']\n",
            "pred:  1 label:  1 ['yeah', 'analysis', 'on', 'cumulative', 'numbers', 'is', 'very', 'hard', 'at', 'national', 'scale', 'because', 'of', 'all', 'state', 'caveats']\n",
            "pred:  0 label:  0 ['mumbai', 'to', 'west', 'bengal', 'migrant', 'train', 'on', '10th', 'may', '2020']\n",
            "pred:  1 label:  1 ['tests', 'per', 'million', 'stand', 'at', 'as', 'of', 'today', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'shows', 'man', 'claiming', 'to', 'be', 'of', 'tablighi', 'jamaat', 'muslim', 'spitting', 'on', 'police', 'when', 'they', 'were', 'putting', 'him', 'into', 'quarantine', '.']\n",
            "pred:  0 label:  0 ['you', 'have', 'had', 'a', 'flu', 'shot', 'in', 'last', '3-5', 'years', 'you', 'will', 'probably', 'test', 'for', '.']\n",
            "pred:  0 label:  0 ['are', 'not', 'some', 'people', 'deserve', 'to', 'be', 'locked', 'up', '.']\n",
            "pred:  0 label:  0 ['joe', 'claim', 'that', 'a', 'different', 'pandemic', 'response', 'from', 'trump', 'would', 'have', 'prevented', 'every', 'coronavirus', 'death', 'goes', 'too', 'far', 'experts', 'said', '.', 'we', 'rated', 'it', 'false', '.', '.']\n",
            "pred:  1 label:  1 ['even', 'without', 'ny', 'historical', 'bump', 'however', 'reported', 'deaths', 'would', 'still', 'have', 'been', 'quite', 'high', 'around', '2000', '.']\n",
            "pred:  0 label:  0 ['an', 'alleged', 'audio', 'of', 'a', 'argentinian', 'prisoner', 'who', 'claims', 'that', '1900', 'inmates', 'will', 'be', 'released', 'from', 'devoto', 'prison', 'as', 'of', 'monday', '.']\n",
            "pred:  0 label:  0 ['in', 'his', 'fox', 'news', 'sunday', 'interview', 'president', 'trump', 'said', 'have', 'one', 'of', 'lowest', 'mortality', 'rates', 'in', 'from', '.', 'chris', 'wallace', 'him', 'in', 'real', 'time', '.', 'now', 'we', 'have', 'too', '.', 'we', 'false', '.', '.']\n",
            "pred:  0 label:  0 ['salt', 'frequently', 'protects', 'from', 'corona', 'infection', 'better', 'than', 'wearing', 'a', 'mask', '.', '\"']\n",
            "pred:  1 label:  1 ['latest', 'updates', 'from', 'there', 'is', 'one', 'new', 'confirmed', 'case', 'of', 'to', 'report', 'in', 'new', 'zealand', '.', 'case', 'is', 'linked', 'to', 'rosewood', 'cluster', 'as', 'a', 'household', 'contact', 'of', 'an', 'earlier', 'case', '.', '.']\n",
            "pred:  1 label:  1 ['yesterday', 'our', 'laboratories', 'completed', 'tests', 'of', 'which', '2022', 'swabs', 'were', 'taken', 'in', 'community', 'and', '454', 'were', 'taken', 'in', 'managed', 'isolation', 'and', 'quarantine', 'facilities', '.', 'that', 'brings', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  0 label:  0 ['man', 'ordered', 'family', 'out', 'on', 'military', 'manouevres', '.']\n",
            "pred:  1 label:  1 ['sir', 'patrick', 'vallance', 'told', 'a', 'downing', 'street', 'briefing', 'that', 'number', 'of', 'new', 'cases', 'was', 'doubling', 'roughly', 'every', 'seven', 'days', '.']\n",
            "pred:  1 label:  1 ['results', 'of', 'genome', 'testing', 'from', 'two', 'auckland', 'cases', 'that', 'were', 'under', 'investigation', 'have', 'returned', 'this', 'morning', 'confirming', 'one', 'case', 'is', 'connected', 'to', 'auckland', 'august', 'cluster', '.']\n",
            "pred:  1 label:  1 ['new', 'study', 'by', 'team', 'published', 'today', 'in', 'shows', 'lockdown', 'and', 'school', 'closures', 'in', 'europe', 'may', 'have', 'prevented', '3', '.', '1', 'million', 'deaths', '.']\n",
            "pred:  0 label:  0 ['ellie', 'goulding', 'says', 'she', 'understand', 'why', 'coronavirus', 'means', 'f1', 'has', 'to', 'stop', '.', 'drivers', 'just', 'keep', 'their', 'helmets', 'on', 'whole', 'time', '?', '”', 'she', 'said', '.']\n",
            "pred:  1 label:  1 ['our', 'total', 'number', 'of', 'confirmed', 'cases', 'to', 'date', 'is', '1387', '.', 'our', 'labs', 'completed', 'tests', 'yesterday', '.', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'is', '.']\n",
            "pred:  0 label:  0 ['graphics', 'with', 'information', 'about', 'containing', 'unicef', 'branding', 'and', 'inaccurate', 'information', '.']\n",
            "pred:  1 label:  1 ['in', 'collaboration', 'with', 'our', 'partners', 'at', 'like', 'to', 'announce', 'beta', 'release', 'of', 'race', 'and', 'ethnicity', 'data', 'for', 'racial', 'data', 'tracker', '.', '.']\n",
            "pred:  1 label:  1 ['give', 'up', 'on', 'patients', 'who', 'still', 'report', 'symptoms', 'months', 'later', '.', '.']\n",
            "pred:  0 label:  0 ['of', 'ayush', ',', 'govt', '.', 'of', 'india', 'suggested', 'use', 'of', 'medicine', 'for', 'its', 'possible', 'role', 'in', 'said', 'dr', '.', 'anil', 'khurana', ',', 'director', 'general', ',', 'central', 'council', 'for', 'research', 'in', 'homoeopathy']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'states', 'reported', 'tests', 'cases', 'and', '1129', 'deaths', '.', 'virginia', 'did', 'not', 'publish', 'new', 'data', 'in', 'time', 'for', 'update', '.', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '.', '5', 'million', 'tests', 'up', 'from', 'yesterday', '.', 'major', 'today', 'california', 'reported', 'tests', 'clearing', 'some', 'backlog', '.', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'for', 'details', '.']\n",
            "pred:  1 label:  1 ['although', 'is', 'a', 'risk', 'for', 'all', 'indigenous', 'peoples', 'globally', 'who', 'is', 'deeply', 'concerned', 'about', 'impact', 'of', 'virus', 'on', 'indigenous', 'peoples', 'in', 'americas', 'which', 'remains', 'current', 'epicenter', 'of']\n",
            "pred:  0 label:  0 ['photo', 'shows', 'bedding', 'and', 'food', 'aid', 'packages', 'distributed', 'in', 'rwanda', 'during', 'coronavirus', 'pandemic']\n",
            "pred:  1 label:  1 ['active', 'cases', 'only', 'of', 'total', 'cases', 'recoveries', 'are', 'now', '26', 'lakh', '.', 'exceed', 'active', 'cases', 'by', 'more', 'than', '18', 'lakh', '.', '.']\n",
            "pred:  1 label:  1 ['a', 'good', 'catch', '.', 'thank', 'you', '.', 'correct', 'for', 'today', 'and', 'work', 'back', 'through', 'a', 'historical', 'scrub', '.', 'we', 'have', '2', 'other', 'screen', 'grabs', 'we', 'can', 'use', 'it', 'looks', 'like', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'purporting', 'that', 'an', 'attack', 'on', 'doctors', 'was', 'avenged', 'by', 'policemen', '.']\n",
            "pred:  0 label:  0 ['two', 'psychics', 'and', 'a', 'book', 'about', 'cia', 'have', 'predicted', 'pandemic', '.']\n",
            "pred:  1 label:  1 ['rt', '!', '6th', 'consecutive', 'day', 'number', 'of', 'new', 'confirmed', 'cases', 'has', 'been', 'lower', 'than', 'new']\n",
            "pred:  0 label:  0 ['a', 'video', 'of', 'a', 'policeman', 'taking', 'down', 'a', 'man', 'wandering', 'during', 'a', 'in', 'italy', '.', '\"']\n",
            "pred:  1 label:  1 ['your', 'risk', 'of', 'increases', 'closer', 'you', 'get', 'to', 'others', 'and', 'longer', 'you', 'interact', 'with', 'them', '.', 'learn', 'ways', 'to', 'help', 'safely', 'resume', 'daily', 'activities', 'and', 'slow', 'spread', 'of', '.', '.', '.']\n",
            "pred:  1 label:  1 ['while', 'reimbursement', 'will', 'go', 'some', 'way', 'to', 'helping', 'leaders', 'with', 'management', 'of', 'crisis', 'figure', 'pales', 'in', 'comparison', 'to', '.', 'officials', 'think', 'was', 'paid', 'out', 'in', 'error', 'or', 'to', 'fraudsters', '.']\n",
            "pred:  1 label:  1 ['according', 'to', '.', 'as', 'of', 'yesterday', 'evening', 'have', 'died', 'of', 'in', 'united', 'states', '.', 'over', 'previous', '24', 'hours', 'that', 'includes', 'deaths', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'gives', 'your', 'weekend', 'safety', 'briefing', '.']\n",
            "pred:  0 label:  0 ['oxford', 'coronavirus', 'vaccine', 'is', 'safe', 'and', 'induces', 'strong', 'immune', 'response', 'early', 'trial', 'results', 'suggest', '.']\n",
            "pred:  0 label:  0 ['everyone', 'should', 'ensure', 'their', 'mouth', 'and', 'throat', 'are', 'moist', ',', 'never', 'dry', '.', 'advice', 'attributed', 'to', 'japanese', 'doctors', 'treating', 'cases', 'further', 'reads', 'that', 'take', 'a', 'few', 'sips', 'of', 'water', 'every', '15', 'minutes', 'at', 'least', '.']\n",
            "pred:  1 label:  1 ['there', 'are', 'currently', 'people', 'in', 'managed', 'isolation', 'and', 'quarantine', '.', 'our', 'current', 'effective', 'capacity', 'is', '.', 'this', 'gives', 'us', 'an', 'excess', 'capacity', 'of', '1932', '.', 'over', 'next', 'week', 'we', 'are', 'projecting', '2606', 'arrivals', 'and', '2810', 'departures', 'from', 'our', 'facilities', '.']\n",
            "pred:  0 label:  0 ['says', 'a', 'photo', 'shows', 'south', 'carolina', 'after', 'stay-at-home', 'orders', 'were', 'announced', '.']\n",
            "pred:  0 label:  0 ['people', 'lying', 'in', 'street', 'in', 'china', 'because', 'of', 'new', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['in', 'april', 'together', 'with', ';', 'multiple', 'other', 'partners', 'who', 'launched', 'access', 'to', 'tools', 'accelerator', 'to', 'catalyse', 'development', 'of', 'and', 'equitable', 'access', 'to', 'vaccines', 'diagnostics', ';']\n",
            "pred:  1 label:  1 ['today', 'ncdc', 'was', 'part', 'of', 'daily', 'press', 'briefing', 'by', 'presidential', 'task', 'force', 'on', 'control', 'led', 'by', '.', 'as', 'at', '1st', 'april', '2020', '151', 'confirmed', 'cases', 'of', 'have', 'been', 'recorded', 'with', 'nine', 'discharged', 'cases', 'and', 'two', 'deaths', '.', '.']\n",
            "pred:  1 label:  1 ['this', 'includes', 'four', 'related', 'to', 'recent', 'case', 'involving', 'a', 'health', 'worker', 'at', 'auckland', 'quarantine', 'facility', '.']\n",
            "pred:  1 label:  1 ['we', 'will', 'provide', 'further', 'advice', 'on', 'where', 'masks', 'can', 'be', 'sourced', '.', 'nz', 'tracer', 'has', 'now', 'recorded', 'more', 'than', 'registered', 'users', '.', 'there', 'have', 'been', 'posters', 'created', 'and', 'poster', 'scans', '.', 'there', 'have', 'been', 'manual', 'entries', 'recorded', 'in', 'app', '.']\n",
            "pred:  0 label:  0 ['facebook', 'posts', 'shared', 'thousands', 'of', 'times', 'accuse', 'authorities', 'of', 'lining', 'streets', 'with', 'empty', 'body', 'bags', 'in', 'a', 'bid', 'to', 'fool', 'people', 'about', 'novel', 'coronavirus', 'pandemic', '.']\n",
            "pred:  0 label:  0 ['total', 'number', 'of', 'deaths', 'in', 'brazil', 'have', 'decreased', 'during', 'pandemic', '.']\n",
            "pred:  0 label:  0 ['mass', 'disinfection', 'of', 'people', 'using', 'a', 'chemical', 'solution', 'will', 'eradicate']\n",
            "pred:  1 label:  1 ['there', 'are', '54', 'people', 'linked', 'to', 'community', 'cluster', 'who', 'remain', 'in', 'auckland', 'quarantine', 'facility', 'which', 'includes', '22', 'people', 'who', 'have', 'tested', 'positive', 'for', 'and', 'their', 'household', 'contacts', '.']\n",
            "pred:  1 label:  1 ['we', 'also', 'just', 'a', 'number', 'of', 'new', 'cases', 'for', 'line', 'with', 'recent', 'growing', 'numbers', '.', 'arizona', 'remains', 'a', 'trouble', 'spot', '.', 'alabama', 'which', 'is', 'experiencing', 'an', 'increase', 'in', 'cases', 'recently', 'began', 'reporting', 'hospitalization', '.', '.']\n",
            "pred:  1 label:  1 ['sadly', 'there', 'are', 'two', 'people', 'with', 'in', 'hospital', 'today', ';', 'one', 'in', 'north', 'shore', 'general', 'ward', 'and', 'one', 'in', 'icu', 'in', 'waikato', 'hospital', '.', 'we', 'wish', 'them', 'a', 'full', 'and', 'fast', 'recovery', '.']\n",
            "pred:  0 label:  0 ['worlds', 'first', 'vaccine', 'in', 'russia', 'given', 'to', 'daughter']\n",
            "pred:  1 label:  1 ['helping', 'patients', 'through', 'care', 'syndrome', '.', '.', '.']\n",
            "pred:  0 label:  0 ['france', 'classifies', 'tunisia', 'as', 'red', 'zone', 'due', 'to', ',', 'while', 'tunisia', 'classifies', 'france', 'as', 'a', 'green', 'area']\n",
            "pred:  0 label:  0 ['israel', 'has', 'manufactured', 'coronavirus', 'vaccine']\n",
            "pred:  0 label:  0 ['bill', '&', 'melinda', 'gates', 'foundation', 'building', 'says', 'for', 'global', 'human', 'population', 'on', 'its', 'front', '.']\n",
            "pred:  1 label:  1 ['states', 'reported', 'more', 'than', 'new', 'cases', 'today', 'which', 'is', 'a', 'higher', 'daily', 'number', 'than', 'seen', 'since', 'mid-may', '.', '.']\n",
            "pred:  1 label:  1 ['health', 'secretary', 'matt', 'hancock', 'has', 'admitted', 'there', 'are', 'with', 'testing', 'system', 'following', 'a', 'in', 'demand', '.', 'more', 'on', 'this', 'story', '.', '.']\n",
            "pred:  0 label:  0 ['bloggers', 'say', 'vatican', 'reports', 'that', 'pope', 'francis', 'tested', 'negative', 'for', 'coronavirus', 'before', 'a', 'second', 'test', 'came', 'back', 'positive', '.']\n",
            "pred:  1 label:  1 ['everyone', 'is', 'expected', 'to', 'make', 'best', 'effort', 'to', 'contain', 'this', '.']\n",
            "pred:  0 label:  0 ['post', 'says', 'will', 'do', 'nothing', 'for', 'coronavirus', '.', '\"']\n",
            "pred:  1 label:  1 ['weekly', 'forecasts', 'of', 'deaths', 'in', 'multiple', 'countries', 'full', 'report', '.', '.']\n",
            "pred:  0 label:  0 ['twelve', 'florida', 'marlin', 'players', 'are', 'stricken', 'with', 'virus', '.']\n",
            "pred:  0 label:  0 ['man', 'on', 'week-long', 'social', 'media', 'break', 'wonders', 'why', 'no', 'one', 'is', 'at', 'this', 'awesome', 'buffet', '.', '.']\n",
            "pred:  0 label:  0 ['kennedy', 'center', ',', 'pbs', 'donated', 'to', 'democrats', 'after', 'getting', 'stimulus', 'money', '.']\n",
            "pred:  1 label:  1 ['other', 'complaints', 'relate', 'to', 'not', 'maintaining', 'proper', 'distancing', 'in', 'amount', 'of', 'gatherings', 'was', 'lower', 'than', 'in', 'previous', 'weekends', ';', 'focus', 'remains', 'on', 'education', 'encouragement', 'and', 'a', 'total', 'of', '30', 'breaches', '29', 'of', 'those', 'resulted', 'in', 'warnings', '1', 'prosecution']\n",
            "pred:  0 label:  0 ['half', 'a', 'million', 'of', 'people', 'are', 'demanding', 'investigation', 'on', 'bill', 'gates', 'charging', 'him', 'with', 'crimes', 'against', 'humanity', 'an', 'investigation', 'from', 'white', 'house', 'is', 'on', 'way', '.']\n",
            "pred:  1 label:  1 ['as', 'many', 'as', 'of', 'people', 'with', 'never', 'realize', 'they', 'had', 'infection', '.', '.']\n",
            "pred:  1 label:  1 ['we', 'want', 'to', 'share', 'a', 'new', 'resource', 'with', 'followers', 'of', 'this', 'a', 'knowledge', 'base', 'for', 'antibody', '.', '.', 'thanks', 'to', 'our', 'friends', 'at', 'for', 'free', 'pro', 'plan', '.', 'your', 'product', 'has', 'been', 'indispensable', 'for', 'building', 'this', 'knowledge']\n",
            "pred:  1 label:  1 ['r', 'number', 'rate', 'of', 'infection', 'of', 'has', 'risen', 'to', 'between', '1', '.', '1', 'and', '1', '.', '4', 'in', 'uk', '.', 'says', 'it', 'helps', 'to', 'explain', 'why', 'we', 'are', 'seeing', 'tighter', 'restrictions', 'across', 'country', '.', 'more', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'and', 'can', 'result', 'in', 'complications', '.', 'however', 'is', 'associated', 'with', 'additional', 'complications', 'like', 'blood']\n",
            "pred:  0 label:  0 ['are', 'kids', 'more', 'likely', 'to', 'suffer', 'from', 'flu', 'than', '?', 'possible', 'but', 'not', 'conclusive', '.', 'incorrectly', 'uses', 'rather', 'than', 'deaths', '.', 'his', 'claim', 'downplays', 'risks', 'by', 'focusing', 'only', 'on', '.']\n",
            "pred:  0 label:  0 ['is', 'a', 'gas', '.', 'this', 'is', 'what', 'happened', 'tonight', 'in', 'streets', 'of', 'rome', 'thousands', 'of', 'birds', 'lying', 'everywhere', '.']\n",
            "pred:  0 label:  0 ['great', 'idea', 'for', 'mass', 'steam', 'inhalation', 'during', 'this', '.']\n",
            "pred:  0 label:  0 ['photo', 'shows', 'muslims', 'in', 'tamil', 'nadu', 'state', 'of', 'india', 'are', 'floating', 'lockdown', 'rules', 'during', 'ramzan', 'to', 'offer', 'night', 'prayers', '.']\n",
            "pred:  1 label:  1 ['india', 'tracker', 'on', '06', 'september', '2020', '08:00', '.', '.', '.', 'via', '.']\n",
            "pred:  0 label:  0 ['a', 'picture', 'which', 'states', 'that', 'wearing', 'a', 'protective', 'mask', 'is', 'a', 'inescapable', 'because', 'it', 'leads', 'to', 'consuming', 'a', 'dangerous', 'amount', 'of', 'carbon', 'dioxide', '.']\n",
            "pred:  0 label:  0 ['tasuku', 'honjo', ',', 'a', 'japanese', 'nobel', 'laureate', 'in', 'medicine', 'is', 'claiming', 'that', 'is', 'a', 'chinese-made', 'virus', '.']\n",
            "pred:  1 label:  1 ['testing', 'status', 'stated', 'that', 'samples', 'tested', 'upto', 'september', '16', '2020', 'sample', 'tested', 'on', 'september', '16', '2020', '.']\n",
            "pred:  0 label:  0 ['workforce', 'in', 'country', 'may', 'have', 'just', 'been', 'by', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['a', 'student', 'from', 'pondicherry', 'university', 'in', 'india', 'has', 'found', 'a', 'home', 'remedy', 'for', 'that', 'has', 'been', 'accepted', 'by', 'world', 'health', 'organization', '.']\n",
            "pred:  0 label:  0 ['minister', 'railway', 'said', 'that', 'they', 'close', '12', 'trains', 'immediately', 'as', 'they', 'dont', 'have', 'money', 'to', 'refund', 'to', 'passengers', 'who', 'have', 'bought', 'tickets', 'in', 'advance', '.', 'in', 'my', 'view', 'money', 'be', 'given', 'preference', 'over', 'lives', ',', 'my', 'request', 'is', 'to', 'review', 'decision', '.']\n",
            "pred:  1 label:  1 ['dr', 'bloomfield', 'encourages', 'anybody', 'with', 'respiratory', 'symptoms', 'to', 'seek', 'advice', 'early', 'from', 'healthline', 'or', 'their', 'gps', '.', 'testing', 'is', 'free', '.', 'even', 'though', 'we', 'are', 'in', 'a', 'favourable', 'position', 'in', 'nz', 'with', '12', 'days', 'with', 'no', 'cases', 'we', 'cannot', 'afford', 'to', 'let', 'our', 'guard', 'down', '.', 'play', 'it', 'safe', 'and', 'be', 'kind', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'trump', 'said', '?', '?', '?', 'hundreds', '?', '?', '?', 'of', 'governors', 'are', 'calling', 'him', '?', '?', '?', '&', 'we', 'only', 'have', '50', '.', '?', '?']\n",
            "pred:  1 label:  1 ['community', 'mitigation', 'actions', 'can', 'push', 'peak', 'later', 'and', 'make', 'it', 'lower', 'than', 'it', 'would', 'have', 'without', 'those', 'interventions', '.', 'learn', '.', '2', 'of', '2', '.']\n",
            "pred:  1 label:  1 ['attend', 'next', 'coca', 'call', 'tuesday', 'sept', '.', '15', 'at', '2pm', 'et', 'via', 'zoom', '.', 'telehealth', ';', 'health', 'considerations', 'for', 'addressing', 'health', 'disparities', 'during', 'pandemic', '.', 'learn', '.', '.', '.']\n",
            "pred:  1 label:  1 ['one', 'last', 'our', 'daily', 'update', 'is', 'out', '.', 'now', 'tracked', '2', '.', '36', 'million', 'tests', 'up', 'from', 'yesterday', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'and', 'not', 'all', 'states', 'report', 'all', 'tests', '.', 'for', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['founder', 'told', 'sky', 'news', 'it', 'could', 'provide', 'capacity', 'to', 'test', 'one', 'million', 'people', 'in', 'uk', 'each', 'day', 'within', 'a', 'month', '.']\n",
            "pred:  0 label:  0 ['gates', 'foundation', 'a', 'polio', 'vax', 'in', 'india', 'between', '2000', '&', '2017', 'and', 'paralysed', '496', ',', '000', 'children', '.', '\"']\n",
            "pred:  0 label:  0 ['inhaling', 'steam', 'of', 'orange', 'peels', 'and', 'sea', 'salt', 'will', 'kill', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['india', 'tests', 'highest', 'ever', 'single', 'day', 'tests', 'at', '.', 'cumulative', 'tests', 'increased', 'to', 'more', 'than', '2', '.', '6', 'crore', '.', 'tests', 'per', 'million', 'for', 'has', 'jumped', 'to', '.']\n",
            "pred:  1 label:  1 ['nearly', 'of', 'new', 'cases', 'are', 'reported', 'from', '5', 'states', '.', 'they', 'have', 'also', 'contributed', 'of', 'new', 'recovered', 'cases', '.', '.']\n",
            "pred:  1 label:  1 ['physical', 'distancing', 'is', 'key', 'to', 'break', 'chain', 'of', 'transmission', '.', 'always', 'keep', 'a', 'safe', 'distance', 'from', 'others', '.', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'found', 'in', 'prince', 'hotel', ',', 'one', 'of', 'worker', 'inside', 'had', 'virus', 'and', 'was', 'tested', 'positive', ',', 'some', 'food', 'cooked', 'there', 'was', 'also', 'taken', 'as', 'samples', 'for', 'testing', 'and', 'it', 'was', 'reported', 'positive', 'for', 'virus', ',']\n",
            "pred:  1 label:  1 ['in', 'this', 'instance', 'these', 'individuals', 'should', 'have', 'been', 'tested', 'prior', 'to', 'leaving', 'managed', 'isolation', 'facility', '.', 'ministry', 'of', 'health', 'has', 'put', 'in', 'place', 'a', 'number', 'of', 'actions', 'to', 'make', 'sure', 'anyone', 'arriving', 'into', 'new', 'zealand', 'does', 'not', 'pose', 'any', 'risk', 'from', '.']\n",
            "pred:  0 label:  0 ['government', 'pledges', 'to', 'continue', 'to', 'do', 'all', 'it', 'can', 'to', 'look', 'like', 'doing', 'all', 'it', 'can', 'to', 'defeat', 'coronavirus', '.', '.']\n",
            "pred:  1 label:  1 ['59', 'hand', 'sanitizers', 'are', 'too', 'toxic', 'to', 'use', 'safely', '.', '.', 'via']\n",
            "pred:  0 label:  0 ['a', 'video', 'showed', 'us', 'president', 'trump', 'had', 'a', 'psychic', 'in', 'white', 'house', 'to', 'pray', 'that', 'pandemic', 'will', 'cease', 'soon', '.']\n",
            "pred:  1 label:  1 ['india', 'has', 'scaled', 'yet', 'another', 'more', 'than', '12', 'lakh', 'samples', 'tested', 'in', 'past', '24', 'hours', '.', 'cumulative', 'tests', 'in', 'country', 'have', 'crossed', 'landmark', 'figure', 'of', '6', 'crore', '.', '.']\n",
            "pred:  1 label:  1 ['also', 'draw', 'your', 'attention', 'to', 'a', 'cluster', 'of', 'southern', 'states', 'that', 'are', 'but', 'have', 'rising', 'case', 'loads', 'especially', 'on', 'a', 'per-capita', 'alabama', 'arkansas', 'georgia', 'mississippi', '.', '.']\n",
            "pred:  0 label:  0 ['man', 'dismayed', 'to', 'find', 'his', 'health', 'insurance', 'cover', 'him', 'for', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['and', 'strip', 'club', 'controversy', 'rocks', 'us', 'with', 'mr', '.', 'fauci', 'cautioning', '.']\n",
            "pred:  1 label:  1 ['221', 'new', 'cases', 'of', ';', 'confirmed', 'discharged', '1010', 'deaths', '.']\n",
            "pred:  0 label:  0 ['during', 'a', 'fox', 'news', 'town', 'hall', 'president', 'trump', 'said', 'china', 'worst', 'year', 'had', 'in', '67', 'before', 'coronavirus', 'pandemic', '.', 'we', 'rated', 'a', 'similar', 'claim', 'mostly', 'false', '.', '.', '.']\n",
            "pred:  1 label:  1 ['just', 'these', '10', 'large', 'states', 'reported', 'tests', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'total', 'confirmed', 'cases', 'cases', 'samples', 'tested']\n",
            "pred:  1 label:  1 ['in', 'metrics', 'cases', 'are', 'one', 'of', 'most', 'complex', 'data', 'points', '.', 'we', 'took', 'a', 'very', 'close', 'look', 'at', 'official', 'guidelines', 'and', 'numbers', 'to', 'date', 'and', 'turned', 'up', 'some', 'interesting', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['testing', 'status', 'stated', 'that', 'samples', 'tested', 'upto', 'september', '20', '2020', 'sample', 'tested', 'on', 'september', '20', '2020', '.']\n",
            "pred:  0 label:  0 ['dr', '.', 'li', 'wenliang', 'chinese', 'whistleblower', 'has', 'proposed', 'that', 'drinking', 'tea', 'can', 'cure', '.']\n",
            "pred:  0 label:  0 ['china', 'has', 'successfully', 'tested', 'an', 'serum', 'made', 'in', 'egypt', 'and', 'presented', 'in', 'beijing', 'by', 'egyptian', 'minister', 'of', 'health', '.']\n",
            "pred:  0 label:  0 ['of', 'lemon', 'in', 'a', 'cup', 'of', 'hot', 'water', 'can', 'save', 'your', 'life', '.', 'hot', 'lemon', 'can', 'kill', 'proliferation', 'novel', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['italian', 'priest', 'died', 'refusing', 'to', 'use', 'respirator', 'as', 'he', 'sacrificed', 'it', 'to', 'a', 'younger', 'person', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'u', '.', 's', '.', 'hospitals', 'are', 'preparing', 'for', '96', 'million', 'coronavirus', 'infections', 'and', 'nearly', 'half', 'a', 'million', 'deaths', ',', 'leaked', 'documents', 'reveal', '.', '?', '?']\n",
            "pred:  1 label:  1 ['india', 'has', 'been', 'at', 'forefront', 'of', 'pioneering', 'testing', 'platforms', 'for', 'ramping', 'up', 'testing', 'capacity', 'for', '.', 'we', 'have', 'crossed', 'six', 'crore', 'mark', 'in', 'conducting', 'tests', 'for', 'detection', 'of', 'with', 'an', 'average', 'testing', 'of', 'more', 'than', '10', 'lakh', 'samples', 'per', 'day', '.']\n",
            "pred:  1 label:  1 ['248', 'new', 'cases', 'of', ';', 'cases', 'of', 'in', 'nigeria', '778', '143', '.']\n",
            "pred:  0 label:  0 ['says', 'joe', 'biden', 'and', 'gretchen', 'whitmer', 'were', 'at', 'detroit', 'athletic', 'club', '?', '?', '?', 'while', 'our', 'children', 'are', 'being', 'mandated', 'to', 'wear', 'a', 'mask', 'while', 'playing', 'outdoor', 'sports', '.', '?', '?']\n",
            "pred:  0 label:  0 ['in', 'hong', 'kong', 'people', 'destroyed', 'a', '5g', 'pole', 'because', 'of', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['there', 'have', 'been', 'no', 'cases', 'of', 'infection', 'in', 'beijing', '.', 'epidemic', 'did', 'not', 'create', 'problems', 'for', 'chinese', 'economy', '.', 'this', 'proves', 'that', 'new', 'coronavirus', 'is', 'a', 'biological', 'weapon', 'of', 'chinese', 'to', 'destroy', 'world', '.']\n",
            "pred:  1 label:  1 ['new', 'recent', 'trends', 'show', 'percentage', 'of', 'overall', 'cases', 'in', 'people', 'ages', '18-22', 'years', 'has', 'increased', 'in', 'us', 'especially', 'in', 'midwest', 'south', 'and', 'northeast', '.', 'for', 'latest', 'demographic', 'data', 'visit', 'data', '.', '.']\n",
            "pred:  0 label:  0 ['rt', 'corona', 'ne', 'kuchh', 'ki', ',', 'aur', 'kuchh', 'ka', '.', '.', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'more', 'than', 'patients', 'cured', 'in', '24', 'hours', '.', '.']\n",
            "pred:  1 label:  1 ['there', 'remains', 'one', 'person', 'receiving', 'care', 'for', ';', 'they', 'are', 'in', 'middlemore', 'and', 'are', 'not', 'in', 'icu', '.', 'yesterday', 'our', 'laboratories', 'completed', 'tests', 'bringing', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  0 label:  0 ['purdue', 'pharma', 'recommends', 'oxycontin', 'for', 'treating', 'coronavirus', 'symptoms', '.']\n",
            "pred:  0 label:  0 ['lack', 'of', 'coronavirus', 'testing', 'in', 'u', '.', 's', '.', 'is', 'a', 'national', 'disgrace', '.', '.']\n",
            "pred:  1 label:  1 ['two', 'states', 'in', 'particular', 'have', 'some', 'concerning', 'hospitalization', 'minnesota', 'and', 'north', 'carolina', '.', '.']\n",
            "pred:  1 label:  1 ['states', 'reported', 'more', 'than', 'new', 'cases', 'in', 'line', 'with', 'recent', 'numbers', '.', 'there', 'were', 'no', 'special', 'circumstances', 'today', '.', 'also', 'a', 'list', 'of', '12', 'states', 'reporting', 'more', 'than', '600', 'new', 'cases', 'today', '.', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '23', '.', '5', 'million', 'tests', 'up', 'from', 'yesterday', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'for', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['in', 'india', 'no', 'case', 'of', 'reinfection', 'has', 'been', 'reported', 'to', 'date', '.', 'however', 'as', 'continues', 'to', 'evolve', 'we', 'are', 'learning', 'more', 'about', 'this', 'new', 'virus', 'every', 'day', '.']\n",
            "pred:  1 label:  1 ['there', 'are', 'some', 'important', 'unknowns', 'in', 'current', 'testing', 'data', 'that', 'working', 'to', 'resolve', 'right', 'now', '.', 'we', 'should', 'have', 'a', 'full', 'report', 'on', 'some', 'new', 'state', 'reporting', 'issues', 'in', 'next', 'day', '.']\n",
            "pred:  0 label:  0 ['kroger', 'corporate', 'says', 'employees', 'can', 'not', 'wear', 'flag', 'masks', 'but', 'they', 'can', 'wear', 'blm']\n",
            "pred:  1 label:  1 ['reported', 'daily', 'death', 'toll', 'was', 'in', 'line', 'with', 'last', 'friday', '.', '.']\n",
            "pred:  0 label:  0 ['34', 'members', 'of', 'family', 'come', 'down', 'with', 'coronavirus', 'after', 'holiday', 'event', '.']\n",
            "pred:  1 label:  1 ['mayor', 'of', 'greater', 'manchester', 'andy', 'burnham', 'says', 'government', 'cannot', 'off', 'of', 'furlough', 'scheme', 'in', 'october', 'as', 'a', '10pm', 'curfew', 'is', 'announced', 'for', 'pubs', 'and', 'restaurants', '.', '.', '.']\n",
            "pred:  0 label:  0 ['an', 'alleged', 'bot', 'system', 'on', 'twitter', 'controlled', 'by', 'right', 'wing', 'uncovered', 'by', 'mistake', 'in', 'spain', '.']\n",
            "pred:  1 label:  1 ['odds', 'of', 'developing', 'severe', 'have', 'been', 'found', 'to', 'be', 'as', 'much', 'as', '7', 'times', 'higher', 'in', 'patients', 'with', '.']\n",
            "pred:  0 label:  0 ['rashes', 'are', 'now', 'responsible', 'for', 'hate', '?', 'skin', 'rash', 'website', 'criticised', 'for', 'lack', 'of', 'examples', '.']\n",
            "pred:  0 label:  0 ['trump', 'says', 'if', 'we', 'stopped', 'testing', 'right', 'now', 'have', 'very', 'few', 'cases', 'if', 'any', '.', 'pants', 'on', '.']\n",
            "pred:  0 label:  0 ['photo', 'shows', 'bill', 'clinton', 'among', 'others', 'not', 'wearing', 'a', 'mask', 'at', 'john', 'lewis', '?', '?', '?', 'funeral', '.']\n",
            "pred:  0 label:  0 ['all', 'elective', 'or', 'non-emergency', 'surgeries', 'are', 'banned', 'to', 'prioritize', 'coronavirus', 'treatments', 'and', 'prevent', 'spread', '.']\n",
            "pred:  0 label:  0 ['all', 'lives', 'matter', '!', 'black', 'white', 'brown', 'muslim', 'dalit', 'rich', 'poor', 'and', 'animals', 'surely', 'humans', 'learnt', 'any', 'lesson', 'from', '.', 'nature', 'has', 'its', 'own', 'way', 'of', 'taking', 'revenge', '.', 'karma', 'will', 'strike', '!', '.']\n",
            "pred:  0 label:  0 ['man', 'decides', 'yoga', 'is', 'not', 'for', 'him', '.']\n",
            "pred:  0 label:  0 ['fennel', 'tea', 'is', 'a', 'cure', 'against', 'new', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['president', 'facemask', 'speaks', 'out', '.']\n",
            "pred:  0 label:  0 ['cdc', 'recommends', 'that', 'only', 'people', 'with', 'symptoms', 'should', 'wear', 'masks', '.']\n",
            "pred:  1 label:  1 ['a', 'couple', 'of', 'states', 'in', 'midwest', 'changed', 'way', 'they', 'report', 'and', 'affected', 'topline', 'numbers', '.', 'minnesota', 'changed', 'to', 'reporting', 'total', 'tests', 'not', 'people', 'tested', '.', 'that', 'led', 'to', 'a', 'jump', 'in', 'our', 'data', 'of', 'tests', '.', 'michigan', 'added', 'probable', 'cases', 'to', 'their', 'total', 'adding', '5k', 'cases', '.']\n",
            "pred:  0 label:  0 ['is', 'this', 'new', 'operandi', '!!', 'currency', 'notes', 'lying', 'unclaimed', 'on', 'road', 'triggered', 'panic', 'in', 'hira', 'nagar', 'area', 'of', 'indore', '-', 'a', 'city', 'which', 'has', 'emerged', 'as', 'one', 'prime', 'hotspots', 'in', 'country', '.']\n",
            "pred:  1 label:  1 ['texas', 'has', 'changed', 'their', 'method', 'of', 'counting', 'deaths', 'and', 'hurricane', 'hanna', 'has', 'hit', 'state', '.', 'this', 'may', 'have', 'caused', 'some', 'backlog', 'which', 'could', 'have', 'influenced', 'large', 'number', 'of', 'deaths', '(313)', 'they', 'reported', 'today', '.', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '.', '5', 'million', 'tests', 'up', 'from', 'yesterday', 'a', 'little', 'below', 'april', 'average', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'and', 'not', 'all', 'states', 'report', 'all', 'tests', '.', 'for', 'details', '.']\n",
            "pred:  1 label:  1 ['government', 'ramped', 'up', 'testing', 'quickly', '-', 'but', 'now', 'system', 'is', 'stuttering', '.']\n",
            "pred:  0 label:  0 ['video', 'clip', 'of', 'a', 'leading', 'opposition', 'politician', 'in', 'india', 'has', 'been', 'viewed', 'thousands', 'of', 'times', 'in', 'multiple', 'facebook', 'and', 'twitter', 'posts', 'alongside', 'a', 'claim', 'that', 'it', 'shows', 'him', 'making', 'a', 'confusing', 'remark', 'about', 'india', '?', '?', '?', 's', 'system', 'for', 'classifying', 'regional', 'infections', 'levels', '.', '_']\n",
            "pred:  0 label:  0 ['a', 'list', 'of', 'purported', 'safety', 'guidelines', 'has', 'been', 'shared', 'in', 'multiple', 'facebook', 'and', 'instagram', 'posts', 'that', 'claim', 'it', 'was', 'issued', 'by', 'indian', 'council', 'of', 'medical', 'research', 'leading', 'research', 'group', 'on', 'novel', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['india', 'sets', 'a', 'new', 'nearly', '9', 'lakh', 'tests', 'in', 'one', 'day', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'shows', 'a', 'man', 'being', 'rescued', 'alive', 'from', 'inside', 'a', 'tomb', '.', 'caption', 'states', 'he', 'was', 'buried', 'alive', 'after', 'being', 'declared', 'dead', 'of', '.']\n",
            "pred:  0 label:  0 ['publications', 'that', 'claim', 'that', 'a', 'brazilian', 'deputy', 'has', 'uncovered', 'great', 'farce', 'of', 'sick', 'and', 'deceased', 'of', 'when', 'discovering', 'an', 'empty', 'field', 'hospital', '.']\n",
            "pred:  0 label:  0 ['some', 'about', 'spraying', 'of', 'alcohol', 'or', 'chlorine', 'over', 'body', 'will', 'not', 'kill', 'viruses', ';', 'drinking', 'alcohol', 'smoking', 'or', 'eating', 'garlic', 'does', 'not', 'protect', 'you', 'from', ';', 'sesame', 'oil', 'does', 'not', 'kill', 'it', '.', 'a', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'new', 'because', 'has', 'it', 'listed', 'as', 'one', 'of', 'viruses', 'it', 'kills', '.', '”']\n",
            "pred:  1 label:  1 ['our', 'total', 'number', 'of', 'confirmed', 'cases', 'is', '1220', 'which', 'is', 'number', 'we', 'report', 'to', 'world', 'health', 'organization', '.', 'there', 'is', 'no', 'one', 'in', 'new', 'zealand', 'receiving', 'care', 'for', '.']\n",
            "pred:  1 label:  1 ['of', 'total', 'cases', 'are', 'reported', 'from', 'only', 'five', 'states', '.', 'these', 'states', 'have', 'also', 'reported', 'maximum', 'new', 'recoveries', 'in', 'past', '24', 'hours', '.', '.']\n",
            "pred:  1 label:  1 ['uk', 'has', 'reported', 'new', 'cases', '-', 'up', 'from', 'on', 'sunday', '.', 'read', 'latest', '.', '.']\n",
            "pred:  1 label:  1 ['some', 'big', 'new', 'numbers', 'came', 'in', 'since', 'our', '3/12', '4pm', 'update', '.', 'thanks', 'to', 'we', 'now', 'have', 'a', 'total', 'tested', 'for', 'new', '2314', 'with', '328', 'positives', '.', 'washington', 'also', 'updated', '.', 'tested', 'people', '.', '426', 'positives', '.', '.']\n",
            "pred:  0 label:  0 ['means', 'of', 'identification', 'of', 'vaccination', 'with', 'artificial']\n",
            "pred:  0 label:  0 ['fall', 'for', 'these', 'hoaxes', '—', 'being', 'of', 'a', 'certain', 'race', 'or', 'religion', 'does', 'not', 'make', 'you', 'less', 'likely', 'to', 'be', 'infected', 'by', 'fact', 'checkers', '.', 'via']\n",
            "pred:  1 label:  1 ['prone', 'positioning', 'for', 'severe', 'has', 'potential', 'benefits', 'and', 'harms', '.', '.']\n",
            "pred:  1 label:  1 ['cases', 'per', 'million', 'population', 'is', 'amongst', 'lowest', 'in', 'world', 'which', 'stands', 'at', 'for', 'india', 'and', 'for', 'secretary', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '27', '.', '1', 'million', 'tests', 'up', 'from', 'yesterday', 'keeping', 'average', 'over', '500k', '.', 'for', 'details', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'viral', 'image', 'hints', 'that', 'flu', 'cases', 'are', 'counted', 'as', 'coronavirus', 'cases', 'because', 'there', 'are', 'about', '24', '.', '500', 'less', 'deaths', 'by', 'flu', 'this', 'year', 'than', 'two', 'years', 'ago', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'patient', 'bites', 'a', 'jugular', 'in', 'usa', '.']\n",
            "pred:  0 label:  0 ['says', 'gov', '.', 'tony', 'evers', 'is', '?', '?', '?', 'pushing', 'firearm', 'confiscation', 'orders', '?', '?']\n",
            "pred:  0 label:  0 ['belarus', 'defeated', 'coronavirus', 'without', 'quarantine', '.']\n",
            "pred:  0 label:  0 ['bill', 'gates', 'is', 'creating', 'a', 'vaccine', 'against', 'which', 'would', 'population', '.']\n",
            "pred:  0 label:  0 ['rock', 'group', 'make', 'come', 'back', 'with', 'remake', 'of', 'my', 'sharona', 'as', 'my', 'corona', '.']\n",
            "pred:  1 label:  1 ['pediatric', 'infection', 'rate', 'is', 'a', 'reminder', 'to', 'address', 'social', 'inequities', 'and', 'take', 'virus', 'more', 'seriously', '.', '.']\n",
            "pred:  1 label:  1 ['there', 'are', '161', 'people', 'linked', 'to', 'community', 'cluster', 'who', 'have', 'been', 'transferred', 'to', 'auckland', 'quarantine', 'facility', 'this', 'includes', '88', 'people', 'who', 'have', 'tested', 'positive', 'for', 'and', 'their', 'household', 'contacts', '.']\n",
            "pred:  0 label:  0 ['mosquito', 'bites', 'can', 'fast', 'track', 'transmission']\n",
            "pred:  1 label:  1 ['a', 'further', '366', 'people', 'have', 'been', 'spoken', 'with', 'and', 'referred', 'for', 'testing', '.', 'there', 'are', '367', 'people', 'who', 'we', 'have', 'repeatedly', 'tried', 'to', 'make', 'contact', 'with', 'including', 'via', 'text', 'and', 'via', 'phone', 'calls', '.']\n",
            "pred:  0 label:  0 ['us', 'to', 'shutdown', 'internet', 'to', 'stop', 'spread', 'of', 'new', 'online', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['update', 'there', 'are', 'four', 'new', 'cases', 'of', 'to', 'report', '.', 'two', 'are', 'community', 'cases', 'linked', 'to', 'auckland', 'august', 'cluster', 'and', 'two', 'are', 'imported', 'cases', 'detected', 'at', 'managed', 'isolation', 'facilities', '.']\n",
            "pred:  0 label:  0 ['mms', 'essential', 'oils', 'are', 'cures', 'for', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['gallery', 'like', 'many', 'arts', 'institutions', 'across', 'country', 'has', 'been', 'hit', 'hard', 'by', 'pandemic', 'and', 'is', 'reportedly', 'considering', 'lay-offs', 'thanks', 'to', 'a', 'drop', 'in', 'revenue', '.']\n",
            "pred:  1 label:  1 ['we', 'just', 'announced', 'an', 'award', 'from', 'u', '.', 's', '.', 'government', 'agency', 'for', 'up', 'to', 'million', 'to', 'accelerate', 'development', 'of', 'our', 'mrna', 'vaccine', 'against', 'novel', 'coronavirus', '.', '.']\n",
            "pred:  0 label:  0 ['once', 'again', 'flu', 'vaccine', 'cannot', 'cause', 'you', 'to', 'return', 'a', 'false', 'positive', 'for', '.', 'this', 'and', 'more', 'in', 'latest', '.', 'believe', 'misinformation', 'on', 'internet', '.', 'watch', 'with', 'instead', '.']\n",
            "pred:  1 label:  1 ['numbers', 'presenting', 'here', 'do', 'not', 'include', 'tests', 'that', 'north', 'carolina', 'pulled', 'out', 'of', 'its', 'numbers', 'today', 'because', 'of', 'a', 'reporting', 'error', '.', 'to', 'more', 'accurately', 'present', 'national', 'trends', 'we', 'set', 'new', 'tests', 'to', '0', 'for', 'today', 'in', 'these', 'charts', '.', '.']\n",
            "pred:  0 label:  0 ['15', 'lakh', 'in', 'each', 'account', '15', 'lakh', 'corona', 'count', '.']\n",
            "pred:  1 label:  1 ['648', 'new', 'cases', 'of', ';', 'confirmed', 'discharged', '860', 'deaths', '.']\n",
            "pred:  1 label:  1 ['rt', 'hot', 'if', 'visiting', 'a', 'cooling', 'center', 'follow', 'these', 'safety', 'tips', 'to', 'protect', 'yourself', 'from', '.']\n",
            "pred:  0 label:  0 ['this', 'video', 'shows', 'dead', 'coronavirus', 'victims', 'amassed', 'in', 'bergamo', 'hospital', '.']\n",
            "pred:  1 label:  1 ['new', 'local', 'restrictions', 'are', 'being', 'introduced', 'in', 'northeast', 'england', '-', 'including', 'a', '10pm', 'curfew', 'for', 'bars', 'and', 'pubs', 'and', 'a', 'ban', 'on', 'people', 'mixing', 'with', 'others', 'outside', 'their', 'household', '.']\n",
            "pred:  1 label:  1 ['what', 'state', 'reported', '.', 'highly', 'likely', 'that', 'they', 'are', 'only', 'testing', 'very', 'sick', 'people', 'now', '.', 'testing', 'criteria', 'heavily', 'influence', 'these', 'rates', '.', 'they', 'could', 'also', 'be', 'falling', 'behind', 'on', 'negatives', 'reporting', '.']\n",
            "pred:  1 label:  1 ['vaccine', 'will', 'cost', 'too', 'much', 'to', 'be', 'accessible', 'for', 'most', 'americans', '.', '.']\n",
            "pred:  1 label:  1 ['india', 'tracker', 'on', '02', 'september', '2020', '08:00', '.', '.', '.', 'via', '.']\n",
            "pred:  0 label:  0 ['chinese', 'government', 'is', 'shooting', 'corona', 'people', ',', 'almost', '45000', 'killed']\n",
            "pred:  1 label:  1 ['as', 'at', '5th', 'april', 'there', 'are', '224', 'confirmed', 'cases', '27', 'discharged', '5', 'deaths', 'for', 'a', 'breakdown', 'of', 'cases', 'by', '.', '115', '45', '20', '9', 'akwa', '5', '4', '9', '4', '6', '2', '2', '1', '1']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'focus', 'is', 'on', 'tests', 'states', 'reported', 'tests', 'lowest', 'figure', 'since', 'july', '8', '.', 'there', 'were', 'reported', 'cases', 'and', '1401', 'deaths', '.', '.']\n",
            "pred:  0 label:  0 ['says', 'anthony', 'fauci', '?', '?', '?', 's', 'statement', 'that', 'coronavirus', 'death', 'rate', 'is', '10', 'times', 'that', 'of', 'seasonal', 'flu', 'is', '?', '?', '?', 'a', 'claim', 'without', 'any', 'scientific', 'basis', '.', '\"']\n",
            "pred:  1 label:  1 ['increasing', 'gap', 'between', 'recovered', ';', 'active', 'cases', 'of', 'india', '04', '2020', 'to', 'september', '14', '.']\n",
            "pred:  1 label:  1 ['there', 'are', '7', 'previously', 'reported', 'cases', 'who', 'are', 'considered', 'to', 'have', 'recovered', 'today', 'bringing', 'total', 'number', 'of', 'active', 'confirmed', 'cases', 'to', '136', 'plus', 'one', 'previously', 'reported', 'probable', 'case', 'remains', 'active', '.', 'of', '137', 'active', 'cases', '20', 'are', 'imported', 'cases', 'in', 'facilities', '.']\n",
            "pred:  0 label:  0 ['face', 'masks', 'cause', 'hypoxia', ';', 'wearing', 'it', 'causes', 'one', 'to', 'inhale', 'too', 'much', 'carbon', 'dioxide', 'which', 'can', 'make', 'you', 'sick', '.']\n",
            "pred:  0 label:  0 ['quotes', 'president', 'donald', 'trump', 'as', 'saying', ',', '?', '?', '?', 'everybody', 'says', 'i', '?', '?', '?', 've', 'done', 'a', 'tremendous', 'job', 'with', '?', '?', '?', 'i', 'think', 'a', 'little', 'gratitude', 'would', 'be', 'nice', '.', 'maybe', 'a', 'big', '?', '?', '?', 'thank', 'you', 'mr', '.', 'president', '?', '?', '?', 'is', 'called', 'for', '.', '?', '?']\n",
            "pred:  1 label:  1 ['rt', 'india', 'sets', 'a', 'new', 'highest', 'single', 'day', 'recoveries', 'of', 'nearly', '9', 'lakh', 'tests', 'conducted', 'in', 'one', 'day']\n",
            "pred:  0 label:  0 ['in', 'december', '2019', 'italian', 'politician', 'beppe', 'grillo', 'wore', 'a', 'protective', 'mask', 'in', 'public', '.', 'some', 'people', 'said', 'that', 'he', 'did', 'that', 'because', 'chinese', 'embassy', 'had', 'warned', 'him', 'about', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['government', 'finalising', 'plans', 'to', 'comprehensively', 'blame', 'eu', 'for', 'coronavirus', '.', '.']\n",
            "pred:  0 label:  0 ['ann', 'coulter', 'points', 'out', 'that', 'if', 'both', 'trump', 'and', 'pence', 'are', 'incapacitated', 'by', 'coronavirus', ',', '.']\n",
            "pred:  0 label:  0 ['video', 'of', 'american', 'president', 'donald', 'trump', 'announcing', 'that', 'roche', 'medical', 'company', 'will', 'launch', 'vaccine', 'for', '.']\n",
            "pred:  1 label:  1 ['last', 'week', 'about', '1/3', 'of', 'us', 'reported', 'deaths', 'originated', 'in', 'long-term', 'care', 'facilities', '.', '.']\n",
            "pred:  0 label:  0 ['people', 'should', 'ignore', 'guidelines', 'to', 'wear', 'masks', 'especially', 'since', 'government', 'advise', 'mask', 'wearing', 'for', '.']\n",
            "pred:  0 label:  0 ['a', 'doctor', 'said', 'that', 'a', 'medicine', 'called', 'interferon', 'can', 'cure', 'new', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['five', 'states', 'account', 'for', 'approx', 'of', 'total', 'deaths', 'in', 'country', 'are', 'maharashtra', 'tamil', 'nadu', 'karnataka', 'delhi', ';', 'andhra', 'pradesh', '.', 'rest', '29', '.', 'are', 'from', 'other', 'secretary', '.']\n",
            "pred:  0 label:  0 ['video', 'shows', 'a', 'family', 'that', 'died', 'from', 'novel', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'that', 'claims', 'to', 'feature', 'one', 'of', 'victims', 'of', 'new', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'tracked', 'a', 'total', 'of', 'completed', 'tests', 'up', 'from', 'yesterday', '.', 'a', 'big', 'jump', 'biggest', '.', 'we', 'can', 'only', 'track', 'what', 'a', 'state', 'reports', 'and', 'not', 'all', 'states', 'report', 'all', 'tests', '.', 'for', '.']\n",
            "pred:  0 label:  0 ['serbia', 'has', 'enough', 'coronavirus', 'tests', '.']\n",
            "pred:  1 label:  1 ['rt', 'when', 'you', 'you', 'help', 'protect', 'those', 'around', 'you', 'from', '.', 'when', 'others', 'wear', 'their', 'mask', 'they', 'help', 'protect']\n",
            "pred:  1 label:  1 ['our', 'total', 'number', 'of', 'confirmed', 'cases', 'to', 'date', 'is', '1304', '.', 'yesterday', 'our', 'labs', 'processed', '.', 'our', 'total', 'number', 'of', 'tests', 'to', 'date', 'is', '.']\n",
            "pred:  0 label:  0 ['in', 'brazil', 'people', 'driving', 'without', 'masks', 'will', 'be', 'given', 'traffic', 'tickets', '.']\n",
            "pred:  0 label:  0 ['good', 'people', 'of', 'twitter', ',', 'and', 'especially', 'those', 'with', 'young', 'children', '.', 'if', 'you', 'or', 'your', 'partner', 'have', 'had', 'symptoms', 'during', 'lockdown', ',', 'which', 'best', 'describes', 'your']\n",
            "pred:  1 label:  1 ['there', 'is', 'a', 'need', 'to', 'ensure', 'access', 'to', 'controlled', 'medicines', 'such', 'as', 'sedatives', 'and', 'analgesics', 'for', 'intubation', 'protocols', 'for', 'treatment', 'of', 'patients', 'with', '.', '.', '.']\n",
            "pred:  0 label:  0 ['maharashtra', 'legislative', 'assembly', 'member', 'jitendra', 'tested', 'positive', 'with', 'coronavirus']\n",
            "pred:  1 label:  1 ['india', 'tracker', 'on', '04', 'september', '2020', '08:00', '.', '.', '.', 'via', '.']\n",
            "pred:  1 label:  1 ['with', 'new', 'cases', 'and', '9', 'additional', 'recovered', 'cases', 'our', 'total', 'number', 'of', 'active', 'cases', 'is', '77', '.', 'of', 'those', '33', 'are', 'imported', 'cases', 'in', 'facilities', 'and', '44', 'are', 'community', 'cases', '.']\n",
            "pred:  1 label:  1 ['twelve', 'new', 'cases', 'of', 'have', 'been', 'reported', 'in', 'nigeria', ';', '9', 'in', 'osun', '2', 'in', 'edo', 'and', '1', 'in', 'ekiti', 'state', 'as', 'at', '12:30', 'pm', '1st', 'april', 'there', 'are', '151', 'confirmed', 'cases', 'of', 'reported', 'in', 'nigeria', '.', 'nine', 'have', 'been', 'discharged', 'with', 'two', 'deaths', '.']\n",
            "pred:  1 label:  1 ['parts', 'of', 'london', 'face', 'to', 'prevent', 'second', 'wave', 'of', 'infections', '.']\n",
            "pred:  1 label:  1 ['if', 'you', 'develop', 'common', 'symptoms', 'of', 'like', 'fever', 'cough', 'or', 'sudden', 'loss', 'of', 'taste', 'or', 'smell', 'please', 'do', 'not', '.', 'call', 'your', 'state', 'hotline', 'immediately', 'for', 'advice', '.', 'do', 'it', 'for', 'yourself', 'and', 'for', 'rest', 'of', 'us', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'can', 'be', 'cured', 'by', 'sniffing', 'clove', 'and', 'camphor', 'and', 'by', 'drinking', 'water', ',', 'virus', 'will', 'go', 'to', 'stomach', 'and', 'acid', 'in', 'stomach', 'will', 'kill', 'virus', '.']\n",
            "pred:  0 label:  0 ['perhaps', 'lonely', 'shut', 'ins', 'that', 'are', 'in', 'demographic', 'most', 'likely', 'to', 'spaz', 'out', 'and', 'demand', 'more', 'lockdowns', 'in', 'response', 'to', 'a', 'virus', 'with', 'a', '99', '.', 'survival', 'rate', '?']\n",
            "pred:  0 label:  0 ['trump', 'sign', 'cares', 'law', 'which', 'contains', 'name', 'before', 'epidemic', 'happened', '.']\n",
            "pred:  0 label:  0 ['rinsing', 'mouth', 'with', 'salt', 'water', 'help', 'with', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['end', 'of', 'day', 'we', 'have', 'confirmed', 'cases', 'in', 'all', '50', 'states', '+', 'dc', 'and', '3', 'territories', 'almost', '6000', 'in', 'total', '.', 'wa', 'and', 'ny', 'are', 'now', 'over', '1000', 'cases', 'each', '.', 'of', 'smaller', 'states', 'louisiana', 'colorado', 'and', 'rhode', 'island', 'look', 'like', 'hot', 'spots', '.', '.']\n",
            "pred:  1 label:  1 ['laboratory', 'network', 'with', '16', 'laboratories', 'in', '15', 'countries', 'is', 'now', 'dedicating', 'of', 'its', 'capacity', 'to', 'testing', 'for', '.', 'hundreds', 'of', 'tests', 'are', 'done', 'each', 'day', 'using', 'polio', 'testing', 'machines', 'in']\n",
            "pred:  0 label:  0 ['italian', 'doctors', 'say', 'coronavirus', 'is', 'bacteria', '.']\n",
            "pred:  1 label:  1 ['cdc', 'offers', 'tips', 'to', 'youth', 'sports', 'organizations', 'on', 'ways', 'to', 'protect', 'players', 'families', ';', 'communities', 'and', 'slow', 'spread', 'of', '.', 'learn', '.', '.', '.']\n",
            "pred:  1 label:  1 ['an', 'update', 'on', 'case', 'reported', 'over', 'weekend', '–', 'man', 'who', 'is', 'a', 'recent', 'returnee', 'from', 'india', 'who', 'completed', 'his', 'stay', 'in', 'managed', 'isolation', 'and', 'has', 'subsequently', 'tested', 'positive', 'for', '.']\n",
            "pred:  0 label:  0 ['and', 'chloroquine', 'are', 'cures', 'for', 'new', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['is', 'transmitted', 'from', 'smoke', 'released', 'during', 'cremation', 'of', 'victims']\n",
            "pred:  1 label:  1 ['there', 'are', 'currently', 'people', 'in', 'managed', 'isolation', 'and', 'quarantine', '.', 'our', 'current', 'effective', 'capacity', 'is', '7103', '.', 'this', 'gives', 'us', 'an', 'excess', 'capacity', 'of', '.', 'over', 'next', 'week', 'we', 'are', 'projecting', 'arrivals', 'and', '2421', 'departures', 'from', 'our', 'facilities', '.']\n",
            "pred:  0 label:  0 ['testing', 'policy', 'best', 'in', 'asia', ',', 'probably', 'in', 'world']\n",
            "pred:  0 label:  0 ['starting', 'on', 'may', '20', 'this', 'year', 'egyptian', 'minister', 'of', 'health', 'has', 'placed', 'herself', 'in', 'measures', 'to', 'isolate', 'herself', 'at', 'home', 'as', 'a', 'result', 'of', 'suspected', 'infection', 'with', 'novel', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['couples', 'in', 'england', 'who', 'are', 'in', 'established', 'will', 'no', 'longer', 'have', 'to', 'socially', 'distance', 'if', 'they', 'do', 'not', 'live', 'together', 'government', 'has', 'confirmed', '.']\n",
            "pred:  1 label:  1 ['one', 'person', 'remains', 'in', 'auckland', 'city', 'hospital', 'in', 'a', 'stable', 'condition', 'on', 'a', 'ward', '.', 'our', 'total', 'number', 'of', 'confirmed', 'cases', 'of', 'remains', 'at', '1178', 'which', 'is', 'number', 'we', 'report', 'to', 'world', 'health', 'organization', '.']\n",
            "pred:  1 label:  1 ['rt', 'in', 'past', '5', 'months', 'more', 'than', '3/4', 'of', 'cases', 'have', 'recovered', 'and', 'less', 'than', '1/4', 'are', 'active', 'now', '.']\n",
            "pred:  1 label:  1 ['india', 'scales', 'another', 'peak', 'of', 'single', 'day', 'recoveries', 'active', 'cases', 'discharged', 'in', 'past', '24', 'hours', '.', 'india', 'has', 'been', 'consistently', 'reporting', 'a', 'very', 'high', 'level', 'of', 'daily', 'recoveries', 'of', ';', '70000', 'since', 'past', '11', 'days', '.', '.', '.']\n",
            "pred:  1 label:  1 ['wear', 'a', 'mask', 'in', 'public', '.', 'stay', 'at', 'least', '6', 'feet', 'away', 'from', 'others', 'who', 'live', 'in', 'your', 'household', '.', 'wash', 'your', 'hands', 'often', '.', 'help', 'slow', 'spread', 'of', '.', 'learn', '.', '.', '.']\n",
            "pred:  0 label:  0 ['indian', 'government', 'snooping', 'messages', 'over', 'social', 'media', 'and', 'all', 'social', 'media', 'platforms', 'will', 'be', 'monitored', 'for', 'messages']\n",
            "pred:  0 label:  0 ['who', 'alerted', 'about', 'a', 'new', 'outbreak', 'of', 'virus', 'which', 'is', 'more', 'lethal', 'than', 'novel', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['in', 'a', 'video', 'seen', 'or', 'shared', 'by', 'millions', 'of', 'people', 'doctors', 'in', 'white', 'coats', 'made', 'a', 'series', 'of', 'false', 'or', 'misleading', 'claims', 'on', '.', '.']\n",
            "pred:  1 label:  1 ['india', 'tracker', 'on', '10', 'september', '2020', '08:00', '.', '.', '.', 'via', '.']\n",
            "pred:  0 label:  0 ['everyone', 'who', 'dies', 'while', 'infected', 'with', 'is', 'registered', 'in', 'official', 'documents', 'as', 'by', 'even', 'if', 'death', 'is', 'unrelated', '.', 'g', '.', 'a', 'car']\n",
            "pred:  1 label:  1 ['yesterday', 'our', 'laboratories', 'processed', 'tests', 'for', 'bringing', 'total', 'to', '.', 'further', 'testing', 'is', 'taking', 'place', 'this', 'weekend', 'across', 'country', 'with', 'pop', 'up', 'testing', 'sites', 'in', 'auckland', '.']\n",
            "pred:  0 label:  0 ['one', 'can', 'cure', 'new', 'coronavirus', 'by', 'inhaling', 'vodka', 'fumes', '.']\n",
            "pred:  0 label:  0 ['an', 'image', 'of', 'a', 'doctor', 'goes', 'viral', 'with', 'claim', 'that', 'dr', '.', 'usman', 'riyaz', 'died', 'while', 'treating', 'coronavirus', 'patients', 'in', 'delhi', 'india', '.']\n",
            "pred:  1 label:  1 ['deaths', 'continue', 'falling', '.', '7-day', 'average', 'just', 'edged', 'under', '700', '.', '.']\n",
            "pred:  1 label:  1 ['cases', 'in', 'india', 'on', 'september', '04', '.', '.', 'active', 'cases', '.', 'deaths', 'total', 'confirmed', 'cases', '=', '.']\n",
            "pred:  0 label:  0 ['raissa', 'soares', 'a', 'doctor', 'in', 'bahia', 'was', 'fired', 'from', 'a', 'state', 'hospital', 'under', 'orders', 'from', 'rui', 'costa', 'state', 'governor', '.', 'she', 'was', 'dismissed', 'she', 'favoured', 'use', 'of', 'against', '.']\n",
            "pred:  0 label:  0 ['dr', '.', 'vk', 'srinivas', 'vice-president', 'of', 'bharath', 'biotech', 'he', 'is', 'taking', 'first', 'injection', 'of', '19', 'vaccine', '.', 'they', 'have', 'confidence', 'in', 'their', 'product']\n",
            "pred:  1 label:  1 ['redfield', 'to', 'face', 'masks', 'are', 'most', 'important', 'powerful', 'public', 'health', 'tool', 'we', 'have', '.', '.', 'via']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '18', '.', '2', 'million', 'tests', 'up', 'from', 'yesterday', 'one', 'of', 'larger', 'days', 'so', 'far', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'for', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['at', 'height', 'of', 'pandemic', 'in', 'new', 'york', 'city', 'first', 'responders', 'faced', 'nearly', 'triple', 'number', 'of', 'calls', 'for', 'cardiac', 'arrest', '.', '.']\n",
            "pred:  1 label:  1 ['today', 'at', '2pm', 'et', 'log', 'in', 'to', 'coca', 'facebook', 'profile', 'or', 'call', 'in', '.', 'coca', 'call', 'guidance', 'for', 'dental', 'settings', 'during', 'response', '.', 'learn', 'more', 'about', 'event', 'at', '.', '.', '.']\n",
            "pred:  1 label:  1 ['as', 'of', 'today', 'there', 'are', 'four', 'states', 'in', 'india', 'namely', ';', 'which', 'had', 'zero', 'case', '.', '.']\n",
            "pred:  0 label:  0 ['electoral', 'commission', 'has', 'postponed', '2020', 'elections', 'to', '2023', 'due', 'to', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['recovery', 'rate', 'soars', 'past', '.', 'total', 'recoveries', 'improved', 'to', 'and', 'exceed', 'active', 'cases', 'by', 'nearly', '10', 'lakh', '.', 'actual', 'caseload', 'of', 'country', 'is', 'active', 'cases', 'which', 'is', 'only', '27', '.', 'of', 'total', 'positive', 'cases', '.']\n",
            "pred:  1 label:  1 ['really', 'we', 'need', 'to', 'move', 'on', 'from', 'says', 'expert', '.', '.', '.']\n",
            "pred:  0 label:  0 ['test', 'came', 'back', 'positive', 'for', 'stupidity', '.']\n",
            "pred:  0 label:  0 ['nsw', ',', 'australia', 'positive', 'chinese', 'woman', 'caught', 'on', 'camera', 'spitting', 'on', 'bananas', 'at', 'a', 'suburban', 'supermarket', '.', 'is', 'this', 'an', 'isolated', 'incident', 'or', 'are', 'chinese', 'nationals', 'and', 'communist', 'party', 'loyalists', 'being', 'paid', 'and', 'instructed', 'to', 'do', 'this', 'by', 'their', 'government', '?', '.']\n",
            "pred:  1 label:  1 ['testing', 'only', 'help', 'people', 'track', 'outbreak', '.', 'it', 'also', 'matters', 'for', 'patient', 'care', 'and', 'one', 'of', 'more', 'troubling', 'trends', 'right', 'now', 'is', 'how', 'long', 'it', 'takes', 'to', 'people', 'through', 'negative', 'test', 'results', '.', 'that', 'causes', 'hospitals', 'to', 'burn', 'through', 'ppe', '.']\n",
            "pred:  0 label:  0 ['man', 'who', 'uses', 'whatsapp', ',', ',', 'google', 'maps', ',', 'snapchat', ',', 'instagram', ',', 'facebook', ';', 'twitter', 'apps', 'has', 'privacy', 'concerns', 'about', 'app', '.']\n",
            "pred:  1 label:  1 ['if', '65', 'and', 'older', 'or', 'have', 'an', 'underlying', 'medical', 'condition', 'learn', 'how', 'you', 'can', 'take', 'steps', 'to', 'protect', 'yourself', 'from', '.', 'learn', '.', '.', '.']\n",
            "pred:  1 label:  1 ['most', 'of', 'these', 'changes', 'had', 'minimal', 'impact', '.', 'mn', 'whose', 'units', 'switched', 'from', 'unique', 'ppl', 'to', 'encounters', 'and', 'nh', 'whose', 'units', 'switched', 'from', 'unique', 'ppl', 'to', 'specimens', 'drove', 'increase', '.', 'these', 'rises', 'are', 'due', 'to', 'inclusion', 'of', 'repeat', 'testing', 'which', 'unique', 'ppl', 'counts', 'do', 'not', 'capture', '.']\n",
            "pred:  0 label:  0 ['eleven', '.', 'how', 'many', 'times', 'broken', 'single-day', 'record', 'for', 'new', 'cases', 'in', 'last', 'month', '.', 'this', 'virus', 'going', 'to', 'like', 'president', 'trump', 'wants', '.', 'surging', '—', 'and', 'we', 'need', 'real', 'leadership', 'from', 'this', 'white', 'house', 'to', 'slow', 'its', 'spread', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'shows', 'a', 'doctor', 'falling', 'down', 'because', 'of', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['a', 'photo', 'of', 'vice', 'president', 'of', 'spain', 'in', 'hospital', 'allegedly', 'visiting', 'his', 'father', 'sickened', 'by', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['we', 'also', 'wanted', 'to', 'call', 'your', 'attention', 'to', 'beta', 'release', 'of', 'racial', 'data', 'tracker', 'in', 'collaboration', 'with', '.', 'actively', 'taking', 'feedback', 'building', 'systems', 'and', 'should', 'have', 'a', 'fuller', 'site', 'soon', '.', '.']\n",
            "pred:  0 label:  0 ['police', 'in', 'new', 'orleans', 'confiscate', 'a', 'pole', 'they', 'say', 'was', 'infested', 'with', 'c-19', 'flu', 'virus', '.']\n",
            "pred:  0 label:  0 ['fauci', 'in', 'his', 'own', 'words', '.', 'as', 'criticizes', 'see', 'full', 'words', 'on', 'china', 'travel', 'masks', 'and', 'threat', 'of', 'virus', '.', '.', '.']\n",
            "pred:  0 label:  0 ['high', 'chicken', 'infected', 'from', 'corona', 'virus', 'found', 'in', 'bangalore', 'today', ',', 'kindly', 'circulate', 'message', 'and', 'avoid', 'consumption', 'of', 'chicken', '.', 'spread', 'to', 'your', 'dear', 'ones', '.']\n",
            "pred:  0 label:  0 ['pcr', 'tests', 'were', 'conducted', 'on', 'a', 'journalist', 'and', 'his', 'crew', 'who', 'attended', 'late', 'minister', 'funeral', 'as', 'they', 'had', 'symptoms', '.']\n",
            "pred:  1 label:  1 ['investigational', 'drugs', 'and', 'therapies', '.', '.', '.']\n",
            "pred:  0 label:  0 ['anyone', 'in', 'mumbai', 'requiring', 'plasma', 'for', 'treatment', ',', 'please', 'contact', 'from', 'a', 'group', 'of', 'people', 'ready', 'to', 'make', 'plasma']\n",
            "pred:  1 label:  1 ['states', 'reported', 'more', 'than', '32k', 'cases', 'today', 'and', 'cases', 'yesterday', 'highest', 'numbers', 'seen', 'since', 'may', '1', '.', 'mississippi', 'updated', 'its', 'numbers', 'in', 'two', 'days', 'so', 'we', 'expect', 'missing', '600-800', 'cases', 'there', 'based', 'on', 'their', 'previous', 'updates', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'average', 'daily', 'tests', 'conducted', 'india', 'has', 'set', 'a', 'record', 'in', 'tests', 'conducted', 'per', 'day', '.', '…']\n",
            "pred:  0 label:  0 ['canadian', 'wife', '.', 'this', 'video', 'is', 'for', 'those', 'who', 'still', 'take', 'it', 'seriously', '.']\n",
            "pred:  1 label:  1 ['all', 'new', 'cases', 'were', 'detected', 'as', 'a', 'result', 'of', 'day', '3', 'or', 'day', '12', 'testing', 'and', 'are', 'now', 'in', 'quarantine', '.']\n",
            "pred:  1 label:  1 ['1/10', 'uk', 'could', 'soon', 'see', 'new', 'cases', 'of', 'every', 'day', 'unless', 'action', 'is', 'taken', 'warns', 'chief', 'scientific', 'adviser', '.', 'click', 'below', 'to', 'find', 'out', 'more', '.']\n",
            "pred:  1 label:  1 ['chairman', 'of', 'british', 'medical', 'association', 'dr', 'tells', 'sky', 'news', 'that', 'are', 'extremely', 'worried', 'that', 'not', 'for', 'a', 'second', 'wave', 'of', '.', '.', '.']\n",
            "pred:  0 label:  0 ['is', 'latin', 'for', 'a', 'sheep', '.', 'starts', 'with', 'a', 'c', ',', 'which', 'also', 'means', 'in', 'ancient', 'language', '.', '19', 'was', 'known', 'as', 'of', 'in', 'ancient', 'times', '.', '”', 'it', 'goes', 'on', 'to', 'draw', 'conclusion', ',', '19', '=', 'a', 'sheep', 'surrender']\n",
            "pred:  1 label:  1 ['recovery', 'rate', 'improves', 'to', '80', '.', 'as', 'on', 'september', '22', '2020', 'improvement', 'in', 'recovery', 'rate', 'since', 'initiation', 'on', 'march', '25', '2020', 'via', '.']\n",
            "pred:  0 label:  0 ['woman', 'moves', 'beyond', 'stage', 'of', 'lockdown', ',', 'breaks', 'ground', 'on', 'mill', 'in', 'apartment', '.', '.']\n",
            "pred:  1 label:  1 ['there', 'are', '13', 'previously', 'reported', 'cases', 'who', 'are', 'considered', 'to', 'have', 'recovered', 'today', '–', 'all', 'community', 'cases', '.', 'with', '14', 'new', 'cases', 'our', 'total', 'number', 'of', 'active', 'cases', 'is', '132', '.', 'of', 'those', '33', 'are', 'imported', 'cases', 'in', 'facilities', 'and', '99', 'are', 'community', 'cases', '.']\n",
            "pred:  0 label:  0 ['false', 'positive', 'result', 'rate', 'in', 'pcr', 'tests', 'is', '50', 'percent', '.']\n",
            "pred:  0 label:  0 ['say', 'televangelist', 'pat', 'robertson', 'said', 'some', 'young', 'people', '?', '?', '?', 'are', 'doing', 'all', 'kinds', 'of', 'unnatural', 'things', 'with', 'their', 'sex', 'organs', '.', 'when', 'people', 'do', 'that', ',', 'they', 'transfer', 'all', 'kinds', 'of', 'chemicals', 'from', 'ladies', '?', '?', '?', 'private', 'parts', 'and', 'that', '?', '?', '?', 's', 'where', 'i', 'think', 'came', 'from', '.', '?', '?']\n",
            "pred:  0 label:  0 ['high', 'death', 'rate', 'in', 'italy', 'is', 'due', 'to', 'people', 'taking', 'ibuprofen', '.']\n",
            "pred:  1 label:  1 ['update', 'from', 'today', 'there', 'are', '2', 'new', 'cases', 'of', 'to', 'report', 'in', 'managed', 'isolation', 'facilities', 'in', 'nz', 'it', 'has', 'been', '75', 'days', 'since', 'last', 'case', 'of', 'was', 'acquired', 'locally', 'from', 'an', 'unknown', 'source', '.', 'our', 'total', 'number', 'of', 'active', 'cases', 'is', '27', '-', 'all', 'are', 'in', 'quarantine']\n",
            "pred:  1 label:  1 ['rt', 'some', 'outbreak', 'reports', 'related', 'to', 'crowded', 'indoor', 'spaces', 'have', 'suggested', 'possibility', 'of', 'aerosol', 'transmission', 'combined']\n",
            "pred:  0 label:  0 ['richest', 'man', 'mukesh', 'wife', 'nita', 'ambani', 'received', 'rs', '.', '500', 'as', 'welfare', 'fund', 'from', 'central', 'government', '.']\n",
            "pred:  1 label:  1 ['government', 'has', 'developed', 'a', 'it', 'plan', 'for', 'responding', 'to', 'new', 'cases', 'of', 'in', 'community', 'should', 'there', 'be', 'a', 'resurgence', 'of', 'cases', 'in', 'nz', '.', 'ongoing', 'elimination', 'will', 'remain', 'as', 'strategy', 'for', 'dealing', 'with', 'to', 'keep', 'kiwis', 'safe', 'and', 'support', 'economic', 'recovery', '.']\n",
            "pred:  0 label:  0 ['hand', 'preventing', 'spread', 'of', 'diseases', 'or', 'broad', 'government', 'conspiracy', 'theory', 'that', 'big', 'soap', 'want', 'you', 'to', 'know', 'about', '?']\n",
            "pred:  1 label:  1 ['22', 'have', 'better', 'tests', 'per', 'million', 'than', 'national', 'average', '.', 'goa', 'delhi', 'andhra', 'pradesh', 'and', 'tamil', 'nadu', 'are', 'reporting', 'maximum', 'number', 'of', 'tests', 'in', 'a', 'day', '.', '.', '.']\n",
            "pred:  0 label:  0 ['about', 'to', 'deliver', 'remarks', 'on', 'coronavirus', 'pandemic', '.', 'tune', 'in', 'to', 'watch', '.']\n",
            "pred:  0 label:  0 ['opening', 'ceremony', 'of', 'london', '2012', 'olympics', 'announced', 'that', 'new', 'coronavirus', 'was', 'coming', '.']\n",
            "pred:  0 label:  0 ['smoke', 'from', 'cremation', 'will', 'spread', 'please', 'do', 'not', 'burn']\n",
            "pred:  1 label:  1 ['we', 'just', 'announced', 'a', 'supply', 'agreement', 'with', 'u', '.', 's', '.', 'government', 'for', 'an', 'initial', '100', 'million', 'doses', 'of', 'our', 'vaccine', 'candidate', 'against', '.', 'read', 'more', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'obama', 'administration', 'made', 'a', 'decision', 'on', 'testing', 'that', 'turned', 'out', 'to', 'be', 'very', 'detrimental', 'to', 'what', 'we', '?', '?', '?', 're', 'doing', '?', '?', 'on', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['images', 'of', 'medical', 'facility', 'shared', 'with', 'claim', 'that', 'indian', 'army', 'set', 'up', 'a', 'hospital', 'in', 'rajasthan', 'during', 'coronavirus', 'pandemic', '.']\n",
            "pred:  0 label:  0 ['a', 'whatsapp', 'message', 'says', 'that', 'coronavirus', 'is', 'being', 'transmitted', 'in', 'wastewater', '.']\n",
            "pred:  1 label:  1 ['rt', '!', 'its', 'focussed', 'strategies', ';', 'effective', 'measures', 'india', 'is', 'reporting']\n",
            "pred:  1 label:  1 ['yesterday', 'our', 'laboratories', 'processed', 'tests', 'bringing', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  0 label:  0 ['a', 'photo', 'has', 'been', 'shared', 'in', 'multiple', 'posts', 'on', 'facebook', 'and', 'twitter', 'alongside', 'a', 'claim', 'it', 'shows', 'bodies', 'of', 'people', 'who', 'died', 'in', 'italy', 'after', 'they', 'became', 'infected', 'with', 'novel', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['showing', 'vaccines', 'for', 'coronavirus', 'have', 'already', 'been', 'discovered', 'were', 'patented', 'and', 'will', 'be', 'readily', 'available', '.', '_']\n",
            "pred:  1 label:  1 ['boris', 'johnson', 'is', 'facing', 'demands', 'from', 'labour', 'to', 'explain', 'his', 'proposal', 'to', 'use', 'army', 'to', 'support', 'police', 'amid', 'new', 'lockdown', 'rules', '.']\n",
            "pred:  1 label:  1 ['one', 'caveat', 'on', 'total', 'test', 'minnesota', 'has', 'updated', 'their', 'data', 'recording', 'which', 'resulted', 'in', 'a', 'reduction', 'of', 'cumulative', 'tests', '.', 'we', 'placed', 'a', 'zero', 'in', 'place', 'of', 'for', 'daily', 'chart', 'above', 'so', 'real', 'trend', 'would', 'be', 'more', 'apparent', '.']\n",
            "pred:  0 label:  0 ['when', 'mosquitoes', 'come', 'out', 'everyone', ',', 'wear', 'bug', 'spray', 'outside', '.', 'i', 'might', 'be', 'wrong', ',', 'but', 'i', 'personally', 'think', 'that', 'if', 'a', 'mosquito', 'sucks', 'blood', 'of', 'a', 'person', 'with', 'and', 'then', 'bites', 'you', ',', 'virus', 'could', 'spread', 'to', 'you', '.', 'im', 'not', 'smart', ',', 'but', 'just', 'do', 'it', 'to', 'be', 'safe', '.']\n",
            "pred:  0 label:  0 ['a', 'person', 'with', 'coronavirus', 'is', 'in', 'andorra', 'on', '.']\n",
            "pred:  0 label:  0 ['man', 'buys', 'vast', 'quantity', 'of', 'drinking', 'water', '.']\n",
            "pred:  1 label:  1 ['texas', 'meanwhile', 'set', 'a', 'new', 'record', 'for', 'number', 'of', 'patients', 'currently', 'hospitalized', 'with', '.', '.']\n",
            "pred:  0 label:  0 ['you', 'can', 'test', 'a', 'quality', 'of', 'reusable', 'mask', 'blowing', 'a', 'lighter', 'as', 'you', 'wear', 'it', '.']\n",
            "pred:  0 label:  0 ['us', 'has', 'developed', 'miracle', 'drug', 'against', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '10', '.', '3', 'million', 'tests', 'up', 'from', 'yesterday', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'for', 'details', '.', '.']\n",
            "pred:  0 label:  0 ['says', 'italy', 'arrested', 'a', 'doctor', '?', '?', '?', 'for', 'intentionally', 'killing', 'over', '3', ',', '000', 'coronavirus', 'patients', '.', '?', '?']\n",
            "pred:  0 label:  0 ['a', 'woman', 'calling', 'herself', 'advocate', 'was', 'found', 'in', 'reliance', 'mart', ',', 'rani', 'bagh', ',', 'delhi', 'flouting', 'social', 'distancing', 'rules', ',', 'without', 'mask', ',', 'touching', 'multiple', 'packets', 'and', 'removed', 'her', 'shirt', 'when', 'security', 'took', 'her', 'out', '.']\n",
            "pred:  0 label:  0 ['15', 'years', 'ago', 'published', 'a', 'study', 'on', 'chloroquine', '.', 'it', 'is', 'effective', 'against']\n",
            "pred:  1 label:  1 ['checkpoints', 'will', 'be', 'set', 'up', 'leading', 'into', 'nyc', 'to', 'enforce', 'a', 'quarantine', 'order', 'for', 'people', 'coming', 'from', 'states', 'with', 'high', 'rates', 'of', '.', 'via', 'last', 'week', '.', '.']\n",
            "pred:  1 label:  1 ['u', '.', 's', '.', 'situation', 'continues', 'to', 'be', 'highly', 'influenced', 'by', 'regional', 'decline', 'in', 'ny', 'metro', 'rest', 'of', 'country', 'shows', 'a', 'different', 'pattern', '.', '.']\n",
            "pred:  0 label:  0 ['finds', 'of', 'americans', 'say', 'they', 'will', 'not', 'drink', 'corona', 'beer', 'because', 'of', 'virus', '.', '\"']\n",
            "pred:  0 label:  0 ['donald', 'trump', 'published', 'on', 'his', 'twitter', 'account', 'that', 'brazilian', 'president', 'is', 'a', 'but', 'that', 'his', 'policies', 'are', 'leading', 'to', 'a', '.']\n",
            "pred:  0 label:  0 ['trump', 'has', 'used', 'word', 'hundreds', 'of', 'times', 'as', 'president', 'most', 'often', 'in', 'reference', 'to', 'mueller', 'report', 'and', 'his', 'recent', 'impeachment', '.', 'recent', 'tv', 'ads', 'have', 'attacked', 'trump', 'for', 'using', 'in', 'context', 'of', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['if', 'indiana', 'continues', 'to', 'hospitals', 'should', 'not', 'become', 'overloaded', '.', 'however', 'if', 'distancing', 'is', 'relaxed', 'there', 'is', 'a', 'potential', 'for', 'over', 'hospitalizations', '-', 'overwhelming', 'system', 'by', 'may', '1', '.', 'learn', 'more', 'at', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'post', 'spreading', 'on', 'whatsapp', 'claims', 'there', 'are', '50', 'hospital', 'guards', 'job', 'at', 'hospital', 'in', 'madrid', '.']\n",
            "pred:  0 label:  0 ['biden', 'is', 'on', 'overdose', 'chloroquine', '.', 'corona', 'stage', '4', 'treatment', '.']\n",
            "pred:  0 label:  0 ['means', 'of', 'identification', 'of', 'vaccination', 'with', 'artificial', 'intelligence']\n",
            "pred:  0 label:  0 ['this', 'is', 'a', 'headline', 'every', 'american', 'should', 'administration', 'ended', 'pandemic', 'early-warning', 'program', 'to', 'detect', 'donald', 'shortsighted', 'actions', 'left', 'our', 'nation', 'ill-prepared', 'to', 'deal', 'with', 'this', 'outbreak', '.', '.']\n",
            "pred:  1 label:  1 ['as', 'of', 'september', '14', 'national', 'forecasts', 'predict', '3000', 'to', '7100', 'new', 'deaths', 'will', 'be', 'reported', 'during', 'week', 'ending', 'october', '10', '.', 'these', 'forecasts', 'predict', 'to', 'total', 'deaths', 'in', 'us', 'by', 'october', '10', '.', '.', '.', '.']\n",
            "pred:  1 label:  1 ['there', 'are', '53', 'people', 'linked', 'to', 'community', 'cluster', 'who', 'remain', 'in', 'auckland', 'quarantine', 'facility', 'which', 'includes', '29', 'people', 'who', 'have', 'tested', 'positive', 'for', 'and', 'their', 'household', 'contacts', '.']\n",
            "pred:  0 label:  0 ['féin', 'have', 'coronavirus', ',', 'pass', 'it', 'varadkar', 'tells', 'media', '.']\n",
            "pred:  1 label:  1 ['commitments', 'include', 'agreements', 'with', '35', 'economies', 'as', 'well', 'as', 'which', 'will', 'procure', 'vaccine', 'doses', 'on', 'behalf', 'of', '27', 'member', 'states', 'plus', 'and', '.', '.']\n",
            "pred:  0 label:  0 ['has', 'developed', 'a', 'vaccine', '.', '_']\n",
            "pred:  0 label:  0 ['christine', 'lagarde', 'admits', 'that', 'something', 'must', 'be', 'done', 'about', 'old', 'people', 'living', 'for', 'too', 'long', '.']\n",
            "pred:  0 label:  0 ['scientists', 'are', 'expressing', 'cautious', 'optimism', 'that', 'a', 'coronavirus', 'vaccine', 'can', 'be', 'ready', 'to', 'go', 'by', 'late', 'spring', 'of', '2021', 'although', 'unclear', 'how', 'much', 'longer', 'it', 'would', 'take', 'to', 'distribute', 'vaccine', 'widely', '.', '.', '.']\n",
            "pred:  1 label:  1 ['as', 'on', '19', 'july', '2020', 'in', 'total', 'positive', 'cases', 'are', 'and', 'active', 'cases', 'are', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'pandemic', 'is', 'impacting', 'fight', 'for', 'an', 'generation', '.', 'and', 'remain', 'dedicated', 'to', 'both']\n",
            "pred:  0 label:  0 ['is', 'a', 'lab', 'made', 'biological', 'weapon', 'deployed', 'to', 'destroy', 'trump', 'economy', '.', 'people', 'will', 'say', 'chinese', 'people', 'died', 'why', 'would', 'they', 'kill', 'their', 'own', 'people', '?', '\"', '.', 'remind', 'people', 'that', 'war', 'is', 'about', 'sacrificing', 'your', 'own', 'people', 'to', 'gain', 'from', 'another', 'people', '.']\n",
            "pred:  0 label:  0 ['of', 'government', 'employees', 'will', 'be', 'deducted', '.']\n",
            "pred:  0 label:  0 ['university', 'of', 'oxford', 'scientists', 'say', 'if', 'their', 'coronavirus', 'vaccine', 'passes', 'clinical', 'trials', 'and', 'gets', 'approved', ',', 'they', 'plan', 'to', 'announce', 'it', 'on', 'day', 'donald', 'trump', 'is', 'voted', 'out', 'of', 'office', 'to', 'make', 'it', 'just', 'a', 'really', 'happy', 'day', '.']\n",
            "pred:  0 label:  0 ['burundi', 'had', '274', 'cases', 'of', 'and', 'only', 'one', 'death', 'on', 'april', '30', '.']\n",
            "pred:  1 label:  1 ['and', 'if', 'we', 'can', 'work', 'together', 'we', 'can', 'ensure', 'that', 'all', 'essential', 'workers', 'are', 'protected', 'and', 'proven', 'treatments', 'like', 'dexamethasone', 'are', 'available', 'to', 'those', 'who', 'need', '.']\n",
            "pred:  0 label:  0 ['says', 'wisconsin', '?', '?', '?', 's', 'in-person', 'election', 'has', 'caused', 'a', '?', '?', '?', 'surge', '?', '?', 'in', 'new', 'coronavirus', 'cases']\n",
            "pred:  0 label:  0 ['barack', 'obama', 'former', 'president', 'of', 'united', 'states', 'said', 'that', 'criminals', 'have', 'rights', '.']\n",
            "pred:  1 label:  1 ['as', 'india', 'records', 'more', 'recoveries', 'than', 'new', 'cases', 'gap', 'between', 'recovered', 'cases', 'and', 'active', 'cases', 'is', 'continuously', 'widening', '.', 'recoveries', 'are', 'more', 'than', '5', 'times', 'active', 'cases', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'healthcare', 'join', 'on', 'thursday', 'at', '8am', 'edt', 'for', 'latest', 'infection', 'and', 'prevention', 'control']\n",
            "pred:  1 label:  1 ['recovery', 'rate', 'improves', 'to', '79', '.', 'as', 'on', 'september', '19', '2020', 'improvement', 'in', 'recovery', 'rate', 'since', 'initiation', 'on', 'march', '25', '2020', 'via', '.']\n",
            "pred:  1 label:  1 ['kia', 'ora', 'timothy', 'most', 'up-to-date', 'locations', 'for', 'our', 'cases', 'are', '.', 'please', 'note', 'everyone', 'is', 'in', 'jet', 'park', 'quarantine', 'facility', 'except', 'women', 'from', 'uk', 'who', 'are', 'isolated', 'at', 'home', 'in', 'wellington', '.']\n",
            "pred:  1 label:  1 ['rt', 'millions', 'of', 'britons', 'are', 'living', 'under', 'restrictions', 'and', 'testing', 'system', 'is', 'coming', 'in', 'for', 'heavy']\n",
            "pred:  1 label:  1 ['rt', 'update', 'coordinated', 'efforts', 'of', ';', 'govts', 'on', 'early', 'detection', 'through', 'aggressive']\n",
            "pred:  1 label:  1 ['testing', 'status', 'stated', 'that', 'samples', 'tested', 'upto', 'september', '21', '2020', 'sample', 'tested', 'on', 'september', '21', '2020', '.']\n",
            "pred:  1 label:  1 ['new', 'cases', 'have', 'been', 'around', 'for', 'last', 'two', 'days', '.', '7-day', 'average', 'continues', 'to', 'fall', 'slightly', '.', '.']\n",
            "pred:  1 label:  1 ['today', 'we', 'have', 'no', 'new', 'cases', 'of', 'to', 'report', '.', 'our', 'total', 'number', 'of', 'confirmed', 'cases', 'is', 'now', '1156', 'which', 'is', 'number', 'we', 'report', 'to', 'world', 'health', 'organization', '.', 'our', 'combined', 'total', 'of', 'confirmed', 'and', 'probable', 'cases', 'is', 'now', '1506', '.']\n",
            "pred:  1 label:  1 ['one', 'piece', 'of', 'ppe', 'saves', 'life', 'mission', 'said', 'rusty', 'a', 'cdc', 'health', 'scientist', 'who', 'led', 'inventory', 'effort', '.', 'cdc', 'shared', '9', 'tons', 'of', 'protective', 'gear', 'with', 'responders', '.', 'read', '.', '.', '.']\n",
            "pred:  0 label:  0 ['european', 'health', 'organization', 'announced', 'that', 'eating', 'beef', 'can', 'cure', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['french', 'health', 'authorities', 'reported', 'new', 'infections', 'on', 'saturday', 'setting', 'a', 'record', 'for', 'highest', 'number', 'of', 'daily', 'new', 'cases', 'since', 'pandemic', 'began', '.', 'number', 'of', 'people', 'who', 'have', 'died', 'from', 'infections', 'was', 'up', 'by', '26', 'read', 'here', '.']\n",
            "pred:  1 label:  1 ['if', 'you', 'have', 'symptoms', ';', 'want', 'to', 'get', 'tested', 'call', 'your', 'medical', 'provider', 'first', '.', 'you', 'can', 'also', 'visit', 'your', 'state', 'or', 'local', 'health', 'website', 'to', 'look', 'for', 'local', 'information', 'on', 'testing', '.', 'learn', '.', '.', '.']\n",
            "pred:  0 label:  0 ['listen', 'to', 'some', 'incriminating', 'revelations', 'on', 'conspiracy', '.', 'unraveling', 'corroboration', 'of', 'chinese', 'collaboration', '?', '.']\n",
            "pred:  1 label:  1 ['result', 'of', 'a', 'positive', 'test', 'is', 'you', 'get', 'money', '.', '”', 'health', 'secretary', 'matt', 'hancock', 'says', 'he', 'is', 'people', 'on', 'low', 'income', 'will', 'do', 'right', 'thing', 'and', 'seek', 'a', 'test', 'if', 'they', 'have', 'symptoms', 'of', '.', 'follow', 'live', '.', '.']\n",
            "pred:  0 label:  0 ['video', 'of', 'west', 'sumatra', 'residents', 'in', 'indonesia', 'throwing', 'off', 'chinese', 'tourists', 'and', 'burning', 'their', 'belongings', 'during', 'pandemic', '.']\n",
            "pred:  1 label:  1 ['our', 'total', 'number', 'of', 'confirmed', 'cases', 'is', '1205', 'which', 'is', 'number', 'we', 'report', 'to', 'world', 'health', 'organization', '.', 'there', 'is', 'no', 'one', 'in', 'new', 'zealand', 'receiving', 'care', 'for', '.']\n",
            "pred:  1 label:  1 ['rt', 'we', 'are', 'steadily', 'progressing', 'towards', 'our', 'goal', 'of', '1', '.', '5', 'lakh', 'functional', 'by', '2022', '.', 'our']\n",
            "pred:  0 label:  0 ['it', 'is', 'an', 'advise', 'to', 'wear', 'surgical', 'masks', 'in', 'reverse', 'to', 'make', 'them', 'fully', 'effective', '.']\n",
            "pred:  1 label:  1 ['intent', 'here', 'is', 'to', 'encourage', 'ourselves', 'to', 'strictly', 'follow', 'at', 'policies', 'prevent', 'overload', 'and', 'prevent', 'added', 'deaths', 'that', 'come', 'along', 'with', 'overload', '.']\n",
            "pred:  1 label:  1 ['case', 'fatality', 'rate', 'further', 'dips', 'to', '2', '.', '.', 'active', 'cases', 'reduced', 'to', 'than', 'recorded', '.', 'recoveries', 'exceed', 'active', 'cases', 'by', 'nearly', '7', 'lakh', '.']\n",
            "pred:  0 label:  0 ['cuba', 'sent', '200', '.', '000', 'doctors', 'across', 'world', 'to', 'help', 'nations', 'fighting', 'against', 'pandemic', '.']\n",
            "pred:  1 label:  1 ['public', 'health', 'need', 'resources', 'for', 'sharing', '?', 'check', 'out', 'one-stop', 'shop', 'for', 'resources', 'that', 'range', 'from', 'youth', 'sports', 'to', '.', '.']\n",
            "pred:  0 label:  0 ['kenya', 'had', 'highest', 'cases', 'of', 'in', 'africa', 'and', 'fifth-highest', 'in', 'world', 'on', 'may', '6', '.']\n",
            "pred:  0 label:  0 ['sir', 'cliff', 'richard', 'says', 'had', 'no', 'choice', 'but', 'to', 'as', 'fans', 'keep', 'trying', 'to', 'hug', 'him', 'in', 'street', '.', 'swear', 'those', 'cunts', 'are', 'trying', 'to', 'kill', 'me', ',', '”', 'he', 'moaned', '.']\n",
            "pred:  1 label:  1 ['as', 'of', 'today', 'states', 'have', 'an', 'elevated', 'risk', 'of', 're-opening', '25', 'states', 'have', 'a', 'moderate', 'risk', 'of', 're-opening', 'and', 'states', 'have', 'a', 'reduced', 'risk', 'of', 're-opening', '.', 'go', 'to', '.', 'to', 'find', 'out', 'more', 'about', 'your', 'state', '.', '.']\n",
            "pred:  0 label:  0 ['state', 'of', 'georgia', 'has', 'become', 'coronavirus', 'guinea', 'pig', '.']\n",
            "pred:  0 label:  0 ['brazilian', 'civil', 'registry', 'data', 'shows', 'that', 'deaths', 'by', 'in', 'country', 'exceed', '25', 'thousand']\n",
            "pred:  1 label:  1 ['as', 'of', 'today', '8', 'states', 'are', 'reporting', 'racial', 'demographics', 'for', 'ct', 'il', 'mi', 'mn', 'nc', 'nj', 'sc', 'va', 'and', '5', 'states', 'are', 'reporting', 'racial', 'demographics', 'for', 'ct', 'il', 'la', 'mn', 'nc']\n",
            "pred:  0 label:  0 ['face', 'masks', 'could', 'be', 'giving', 'people', 'immunity', ',', 'researchers', 'suggest']\n",
            "pred:  1 label:  1 ['note', 'also', 'that', 'line', 'we', 'have', 'on', 'previous', 'charts', 'is', 'a', 'minimum', 'estimate', '.', 'itself', 'has', 'upped', 'its', 'estimate', 'of', 'tests', 'needed', 'to', 'safely', 'reopen', 'to', '.', 'other', 'plans', 'have', 'even', 'higher', 'recommendations', '.', '.', '.']\n",
            "pred:  1 label:  1 ['two', 'cases', 'in', 'christchurch', 'will', 'be', 'placed', 'in', 'quarantine', 'section', 'within', 'one', 'of', 'managed', 'isolation', 'facilities', '.', 'been', 'previously', 'assessed', 'as', 'a', 'dual', 'use', 'facility', '–', 'it', 'has', 'an', 'area', 'for', 'high', 'level', 'of', 'clinical', 'care', 'that', 'someone', 'with', 'may', 'require', '.']\n",
            "pred:  0 label:  0 ['cat', 'purposely', 'caught', 'in', 'hope', 'of', 'killing', 'off', 'owner', '.']\n",
            "pred:  1 label:  1 ['of', 'total', 'cases', 'are', 'recorded', 'only', 'in', '5', 'states', '.', 'of', 'total', 'new', 'cases', 'which', 'have', 'been', 'reported', 'in', 'last', '24', 'hours', 'maharashtra', 'has', 'alone', 'contributed', 'more', 'than', '20000', 'and', 'andhra', 'pradesh', 'has', 'contributed', 'more', 'than', '10000', '.', '.']\n",
            "pred:  1 label:  1 ['over', 'last', '24', 'hours', 'ncdc', 'has', 'worked', 'with', 'and', 'port', 'health', 'services', 'in', 'lagos', ';', 'abuja', 'to', 'receive', '269', 'nigerians', 'from', 'india', 'we', 'provided', 'groups', 'with', 'guidance', 'and', 'materials', 'for', 'observation', 'of', 'mandatory', '14-day', '.']\n",
            "pred:  1 label:  1 ['since', 'last', 'sunday', 'of', 'reported', 'tests', 'in', 'arizona', 'have', 'come', 'back', 'positive', '.', 'right', 'next', 'door', 'in', 'new', 'mexico', 'only', '2', '.', 'of', 'tests', 'are', 'coming', 'back', 'positive', '.', '.']\n",
            "pred:  0 label:  0 ['maduro', 'has', 'supported', 'use', 'of', 'herbal', 'infusions', 'to', 'cure', '.']\n",
            "pred:  1 label:  1 ['rt', '?', 'when', 'you', 'it', 'should', 'cover', 'your', 'and', 'to', 'help', 'protect', 'others', 'and', 'slow', 'spread', 'of', '.']\n",
            "pred:  1 label:  1 ['rt', 'mha', 'has', 'issued', 'new', 'guidelines', 'for', 'opening', 'up', 'of', 'more', 'activities', 'in', 'areas', 'outside', '.', 'in']\n",
            "pred:  0 label:  0 ['ag', 'barr', 'suggests', 'an', 'end', 'to', 'coronavirus', 'lockdown', '.']\n",
            "pred:  0 label:  0 ['suggestion', 'that', 'trump', 'urged', 'sick', 'people', 'to', 'get', 'out', 'and', 'vote', 'during', 'pandemic', '.']\n",
            "pred:  0 label:  0 ['a', 'man', 'who', 'attended', 'tablighi', 'jamaat', 'event', 'in', 'delhi', ',', 'india', ',', 'spit', 'on', 'cops', 'intending', 'to', 'spread', 'corona', 'virus', '.']\n",
            "pred:  0 label:  0 ['workers', 'who', 'worked', 'from', '1990', 'to', '2020', 'have', 'right', 'to', 'receive', 'corona', 'relief', 'of', ',', '000', 'from', 'ministry', 'of', 'labour', 'and', 'employment', '.']\n",
            "pred:  1 label:  1 ['some', 'people', 'in', 'us', 'will', 'be', 'at', 'increased', 'risk', 'of', 'depending', 'on', 'their', 'exposures', '.', 'greatest', 'risk', 'is', 'to', 'those', 'who', 'are', 'in', 'close', 'contact', 'with', 'people', 'with', '.', 'people', 'with', 'suspected', 'or', 'confirmed', 'exposure', 'should', 'reach', 'out', 'to', 'their', 'healthcare', 'provider', '.']\n",
            "pred:  0 label:  0 ['claim', 'that', 'football', 'star', 'cristiano', 'ronaldo', 'has', 'converted', 'his', 'hotel', 'chain', 'into', 'coronavirus', 'hospitals', '.']\n",
            "pred:  1 label:  1 ['number', 'of', 'britons', 'who', 'approve', 'of', 'handling', 'of', 'pandemic', 'has', 'fallen', 'to', 'its', 'lowest', 'ever', 'level', 'according', 'to', 'a', 'new', 'poll', 'from', 'yougov', '.', 'follow', 'latest', 'updates', '.']\n",
            "pred:  1 label:  1 ['states', 'reported', 'new', 'cases', 'today', '.', '7-day', 'average', 'is', 'now', 'back', 'where', 'it', 'was', 'in', 'late', 'april', '.', 'positive', 'rate', 'is', 'about', 'half', 'what', 'it', 'was', 'back', 'today', 'was', 'highest', 'that', 'been', 'since', 'may', '16', '.', '.']\n",
            "pred:  0 label:  0 ['trudeau', 'airs', 'a', 'coronavirus', 'update', 'rerun', ',', 'to', 'see', 'if', 'anyone', 'notices', '.', '.']\n",
            "pred:  1 label:  1 ['maharashtra', 'karnataka', 'andhra', 'pradesh', 'uttar', 'pradesh', 'and', 'tamil', 'nadu', 'contribute', '60', '.', 'of', 'total', 'active', 'cases', 'and', 'are', 'also', 'reporting', 'close', 'to', '.', 'of', 'total', 'recovered', 'cases', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'coronavirus', 'hoax', 'fake', 'virus', 'pandemic', 'fabricated', 'to', 'cover-up', 'global', 'outbreak', 'of', '5g', 'syndrome', '.', '?', '?']\n",
            "pred:  0 label:  0 ['minister', 'of', 'health', 'claims', 'that', 'only', 'one', 'case', 'of', 'registered', 'in', 'lithuania']\n",
            "pred:  0 label:  0 ['season', 'is', 'here', '.', 'will', 'rise', 'again', '.', 'but', 'what', 'about', '?', 'can', 'mosquitoes', 'spread', '?', 'we', '.']\n",
            "pred:  1 label:  1 ['12', 'have', 'recovery', 'rate', 'more', 'than', 'national', 'average', '.', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'nobody', 'is', 'found', 'dead', 'of', 'corona', 'in', 'their', 'homes', '.', '.', 'they', 'all', 'die', 'at', 'hospital', '.', '?', '?']\n",
            "pred:  0 label:  0 ['our', 'disgusting', 'sticky', 'tables', 'catch', 'and', 'trap', 'coronavirus', 'safely', ',', 'insists', 'wetherspoons', 'boss', '.', '.']\n",
            "pred:  0 label:  0 ['dr', '.', 'vk', 'srinivas', 'vice-president', 'of', 'bharath', 'biotech', 'he', 'is', 'taking', 'first', 'injection', 'of', '19', 'vaccine', '.', 'they', 'have', 'confidence', 'in', 'their', 'product', '.']\n",
            "pred:  0 label:  0 ['india', 'is', 'suing', 'microsoft', 'founder', 'bill', 'gates', 'because', 'of', 'his', 'vaccine', 'against', 'coronavirus', 'that', 'killed', 'girls', '.']\n",
            "pred:  0 label:  0 ['a', 'very', 'real', 'sense', ',', 'flattened', 'curve', '.', '.', '.', '.', 'number', 'of', 'cases', 'in', 'oklahoma', '?', '?', '\"', 'declined', 'precipitously', '.', '\"']\n",
            "pred:  1 label:  1 ['police', 'have', 'shut', 'down', 'a', 'series', 'of', 'illegal', 'parties', 'overnight', 'as', 'people', 'enjoyed', 'a', 'final', 'weekend', 'of', 'revelry', 'before', 'tougher', 'restrictions', 'come', 'into', 'force', '.']\n",
            "pred:  0 label:  0 ['a', 'link', 'offering', 'registration', 'for', 'united', 'states', 'humanitarian', 'aid', 'to', 'other', 'countries', 'for', 'coronavirus']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'first', 'volunteer', 'in', 'uk', 'coronavirus', 'vaccine', 'trial', 'has', 'died', '.', '?', '?']\n",
            "pred:  0 label:  0 ['every', 'step', 'of', 'way', ',', 'president', 'trump', 'has', 'ignored', 'experts', ',', 'downplayed', 'threat', 'posed', ',', 'and', 'misled', 'american', 'people', '—', 'and', 'all', 'paying', 'price', 'every', 'single', 'day', '.', 'i', 'promise', 'you', ',', 'i', 'will', 'do', 'opposite', '.', '.']\n",
            "pred:  0 label:  0 ['justin', 'trudeau', 'promises', 'to', 'pay', 'off', 'tamil', 'nadu', 'debt', '.']\n",
            "pred:  1 label:  1 ['update', 'from', 'there', 'are', 'three', 'new', 'cases', 'of', 'to', 'report', 'in', 'managed', 'isolation', 'in', 'new', 'zealand', 'today', '.', 'number', 'of', 'active', 'cases', 'in', 'new', 'zealand', 'is', '25', '.', 'it', 'has', 'been', '79', 'days', 'since', 'last', 'case', 'of', 'was', 'acquired', 'locally', 'from', 'an', 'unknown', 'source', '.']\n",
            "pred:  1 label:  1 ['latest', 'update', 'from', 'ministry', 'of', 'today', 'there', 'are', '2', 'cases', 'of', 'to', 'report', 'in', 'nz', '.', 'both', 'have', 'recently', 'returned', 'from', 'overseas', 'and', 'are', 'in', 'managed', 'isolation', 'facilities', '.', 'total', 'number', 'of', 'active', 'cases', 'in', 'nz', 'is', '16', '.', 'there', 'are', 'no', 'cases', 'in', 'community', '.']\n",
            "pred:  0 label:  0 ['brazilian', 'mainstream', 'media', 'did', 'not', 'say', 'anything', 'after', 'former', 'president', 'lula', 'said', 'nature', 'created', 'this', 'monster', 'called', '.']\n",
            "pred:  1 label:  1 ['but', 'unlike', 'our', 'other', 'numbers', 'which', 'can', 'be', 'considered', 'for', 'number', 'of', 'people', 'tested', 'we', 'cannot', 'say', 'that', 'about', 'number', 'of', 'negative', 'test', 'results', '.']\n",
            "pred:  0 label:  0 ['19', 'kits', 'for', 'home', 'linked', 'to', 'tata', 'health']\n",
            "pred:  1 label:  1 ['rt', 'if', 'and', 'when', 'we', 'have', 'an', 'effective', 'vaccine', 'we', 'must', 'also', 'use', 'it', 'effectively', '.', 'i', 'will', 'repeat', 'vaccine']\n",
            "pred:  1 label:  1 ['nearly', 'of', 'total', 'cases', 'are', 'being', 'contributed', 'by', 'five', 'states', 'viz', '.', 'maharashtra', 'andhra', 'pradesh', 'tamil', 'nadu', 'karnataka', 'and', 'uttar', 'pradesh', '.', '.']\n",
            "pred:  0 label:  0 ['by', 'time', 'this', 'pandemic', 'is', 'over', ',', 'india', 'would', 'likely', 'be', 'worst', 'impacted', 'country', 'in', 'world', '.', 'not', 'just', 'in', 'numbers', 'alone', ',', 'but', 'with', 'a', 'shattered', 'ambition', 'of', 'becoming', 'an', 'economic']\n",
            "pred:  0 label:  0 ['people', 'in', 'wuhan', 'are', 'locked', 'inside', 'their', 'houses', 'by', 'chinese', 'government', '.']\n",
            "pred:  0 label:  0 ['shopkeepers', 'sleeping', 'inside', 'shops', 'due', 'to', 'modi', 'handling', 'of']\n",
            "pred:  1 label:  1 ['is', 'a', 'wildfire', 'not', 'a', 'wave', '.', '.', 'via', '.']\n",
            "pred:  1 label:  1 ['india', 'scales', 'yet', 'another', 'records', 'all-time', 'high', 'of', 'daily', 'testing', '.', 'for', 'first', 'time', 'more', 'than', '12', 'lakh', 'tests', 'conducted', 'in', 'last', '24', 'hrs', '.', '.', '.']\n",
            "pred:  1 label:  1 ['on', 'numbers', '.', 'states', 'reported', 'cases', 'second-highest', 'number', 'ever', 'after', 'yesterday', '.', 'average', 'number', 'of', 'daily', 'deaths', 'reported', 'by', 'states', 'over', 'last', '7', 'days', 'rose', 'by', '150', 'from', 'saturday', 'to', 'saturday', '.', '.']\n",
            "pred:  0 label:  0 ['korea', 'finished', 'developing', '10', 'minute', 'diagnostic', 'kit', 'and', 'is', 'now', 'ramping', 'up', 'production', '.', 'they', 'plan', 'to', 'export', '300', '.', '000', 'per', 'week', '-']\n",
            "pred:  0 label:  0 ['raisin', 'prevents', 'clotting', 'thus', 'helps', 'to', 'fight']\n",
            "pred:  1 label:  1 ['there', 'is', 'currently', 'not', 'enough', 'evidence', 'to', 'suggest', 'survivors', 'become', 'immune', 'after', 'recovery', '.', 'physical', 'distancing', 'your', 'hands', 'frequently', 'with', 'soap', ';', 'water', 'to', 'protect', 'yourself', 'and', 'others', '.', '.']\n",
            "pred:  0 label:  0 ['lame', 'coronavirus', 'party', 'infects', '0', 'people', '.']\n",
            "pred:  1 label:  1 ['concept', 'of', 'asymptomatic', 'was', 'at', 'center', 'of', 'confusion', 'and', 'controversy', 'this', 'week', '.', '.', '.']\n",
            "pred:  1 label:  1 ['there', 'is', 'no', 'one', 'in', 'new', 'zealand', 'receiving', 'care', 'for', '.', 'yesterday', 'new', 'laboratories', 'completed', 'tests', 'bringing', 'total', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  0 label:  0 ['researchers', 'developed', 'a', 'drug', 'that', 'can', 'cure', 'in', '3', 'days', '.']\n",
            "pred:  1 label:  1 ['states', 'reported', 'over', 'new', 'cases', 'today', '.', 'within', 'range', 'seen', 'over', 'last', 'month', '.', 'here', 'are', 'states', 'that', 'reported', 'over', '400', 'new', 'cases', '.', 'arkansas', 'arizona', 'and', 'north', 'carolina', 'reported', 'new', 'one-day', 'highs', '.', '.']\n",
            "pred:  0 label:  0 ['india', 'records', 'over', 'new', 'infections', 'taking', 'confirmed', 'case', 'tally', 'to', '20', '.', '8', 'lakh', '.', 'whereas', 'tamil', 'nadu', 'government', 'announces', 'reopening', 'of', 'small', 'temples', 'mosques', 'churches', 'and', 'dargahs', 'from', '10', 'august', '.', '.']\n",
            "pred:  0 label:  0 ['trump', 'suggests', 'donkey', 'penises', 'cure', '.']\n",
            "pred:  0 label:  0 ['air', 'canada', 'promises', '100', 'bonus', 'aeroplan', 'miles', 'to', 'any', 'passenger', 'that', 'gets', '.', '.']\n",
            "pred:  1 label:  1 ['number', 'of', 'people', 'in', 'uk', 'who', 'have', 'tested', 'positive', 'for', 'has', 'increased', 'by', '3330', 'official', 'figures', 'show', '.']\n",
            "pred:  0 label:  0 ['anyone', 'planning', 'to', 'celebrate', 'world', 'steam', 'week', '?', '.', 'its', 'a', 'fake', 'news', '.', 'steam', 'kill', 'coronavirus', 'or', 'cure', 'you', 'from', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'somehow', 'ny', 'ma', 'nj', 'and', 'ct', 'are', 'only', 'safe', 'places', 'to', 'be', 'in', 'us', 'right', 'now', '.', '.']\n",
            "pred:  1 label:  1 ['update', 'today', 'there', 'are', 'five', 'new', 'cases', 'of', 'to', 'report', 'in', 'new', 'two', 'in', 'and', 'three', 'in', 'community', '.', 'this', 'brings', 'our', 'confirmed', 'cases', 'of', 'is', 'now', '1406', 'which', 'is', 'number', 'we', 'will', 'report', 'to', 'world', 'health', 'organization', '.']\n",
            "pred:  0 label:  0 ['parliament', 'calls', 'for', 'arrest', 'of', 'bill']\n",
            "pred:  0 label:  0 ['.', '.']\n",
            "pred:  1 label:  1 ['health', 'minister', 'says', 'more', 'local', 'lockdowns', 'are', 'a', 'possibility', 'in', 'wales', '.', '.']\n",
            "pred:  1 label:  1 ['schools', 'may', 'close', 'again', 'due', 'to', 'lack', 'of', 'tests', 'headteacher', 'warns', '.']\n",
            "pred:  1 label:  1 ['india', 'records', 'highest', 'ever', 'single', 'day', 'recoveries', 'of', 'today', '.', 'recoveries', 'exceed', 'active', 'cases', 'by', 'more', 'than', '9', 'lakh', '.', 'case', 'fatality', 'rate', 'touches', 'a', 'new', 'low', 'of', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'mainstream', 'media', 'pretended', 'there', 'was', 'a', 'deadly', 'surge', 'in', 'cases', 'thanks', 'to', 'wisconsinites', 'voting', '.', '?', '?']\n",
            "pred:  0 label:  0 ['outbreak', 'of', 'idiocy', 'spreading', '10', ',', '000', 'times', 'faster', 'than', 'coronavirus']\n",
            "pred:  0 label:  0 ['use', 'ozone', 'to', 'fight', 'against', 'spread', 'of', 'corona', 'virus', '.']\n",
            "pred:  1 label:  1 ['kia', 'ora', 'ben', 'info', 'about', 'range', 'of', 'measures', 'at', 'every', 'alert', 'level', 'can', 'be', 'found', '.']\n",
            "pred:  1 label:  1 ['recovery', 'rate', 'improves', 'to', '77', '.', 'as', 'on', 'september', '07', '2020', 'improvement', 'in', 'recovery', 'rate', 'since', 'initiation', 'on', 'march', '25', '2020', 'via', '.']\n",
            "pred:  0 label:  0 ['from', 'random', 'posts', 'to', 'renowned', 'gurus', '.', 'they', 'keep', 'making', 'this', 'self', 'test', 'recommendation', 'for', '.', 'we', '.']\n",
            "pred:  0 label:  0 ['unbelievable', 'president', '.', '.']\n",
            "pred:  1 label:  1 ['of', 'uk', 'physicians', 'who', 'needed', 'a', 'test', 'get', 'one', '.', '.', '.']\n",
            "pred:  1 label:  1 ['weekly', 'deaths', 'registered', 'weekly', 'in', 'england', 'and', 'wales', 'of', 'deaths', 'in', 'week', 'ending', '19', 'june', 'are', 'within', 'range', 'we', 'would', 'have', 'expected', 'based', 'on', 'trend', 'analysis', 'of', 'past', '10', 'years', '.', '.']\n",
            "pred:  1 label:  1 ['we', 'are', 'beginning', 'to', 'track', 'antigen', 'testing', 'which', 'appears', 'to', 'be', 'underreported', '.', 'kentucky', 'for', 'instance', 'added', 'over', '10k', 'antigen', 'tests', 'today', 'alone', '.', 'our', 'api', 'now', 'breaks', 'out', 'antigen', 'and', 'pcr', 'testing', 'whenever', 'possible', '.']\n",
            "pred:  0 label:  0 ['biden', 'said', 'cops', 'have', 'died', 'from', 'this', 'year', 'than', 'have', 'been', 'killed', 'on', 'patrol', '.', '”', 'mostly', 'true', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'we', 'can', 'confirm', 'there', 'are', 'two', 'new', 'cases', 'of', 'in', 'new', 'zealand', 'today', 'both', 'linked', 'to', 'recent', 'travel', 'from', 'uk', '.', 'there']\n",
            "pred:  0 label:  0 ['viral', 'text', 'message', 'claims', 'that', 'chemical', 'substances', 'required', 'to', 'cure', 'can', 'be', 'found', 'in', 'tea', '.', 'dr', '.', 'li', 'wenliang', 'had', 'found', 'this', 'cure', 'while', 'researching', 'about', 'coronavirus', 'before', 'his', 'death', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'continues', 'to', 'mutate', 'in', 'minds', '.']\n",
            "pred:  1 label:  1 ['125', 'new', 'cases', 'of', ';', 'cross', 'confirmed', 'discharged', '1048', 'deaths', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'has', 'been', 'viewed', 'thousands', 'of', 'times', 'in', 'facebook', 'posts', 'alongside', 'a', 'claim', 'it', 'shows', 'a', 'sri', 'lankan', 'doctor', 'who', 'invented', 'a', 'rapid', 'test', 'kit', 'for', 'novel', 'coronavirus', 'which', 'causes', 'disease', '.']\n",
            "pred:  0 label:  0 ['carona', 'virus', 'vaccine', 'ready', '.', 'able', 'to', 'cure', 'patient', 'within', '3', 'hours', 'after', 'injection', '.', 'hats', 'off', 'to', 'us', 'scientists', '.', 'trump', 'announced', 'that', 'roche', 'medical', 'company', 'will', 'launch', 'vaccine', 'next', 'sunday', ',', 'and', 'millions', '.', '?']\n",
            "pred:  0 label:  0 ['wearing', 'masks', 'for', 'a', 'long', 'period', 'of', 'time', 'can', 'cause', 'hypoxia', '.']\n",
            "pred:  0 label:  0 ['indian', 'state', 'of', 'kerala', 'opposition', 'leader', 'writes', 'to', 'chief', 'minister', 'vijayan', 'to', 'adopt', 'methods', 'to', 'save', 'only', 'those', 'with', 'stronger', 'immunity', 'from', 'infection', '.']\n",
            "pred:  0 label:  0 ['basically', 'impossible', 'to', 'get', 'any', 'kind', 'of', 'compensation', 'if', 'by', 'a', 'vaccine', '.']\n",
            "pred:  0 label:  0 ['.', 'says', 'us', 'is', 'at', 'top', 'globally', 'in', 'fight', '.', 'it', '.', '.', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'people', 'defying', 'stay-at-home', 'order', 'fill', 'arizona', 'beaches', '.', '?', '?']\n",
            "pred:  1 label:  1 ['people', 'in', 'scotland', 'have', 'been', 'banned', 'from', 'visiting', 'other', 'households', 'indoors', 'in', 'tough', 'new', 'restrictions', 'that', 'go', 'further', 'than', 'those', 'announced', 'for', 'england', '.', 'get', 'more', 'on', 'new', 'measures', 'in', 'scotland', '.', '.']\n",
            "pred:  0 label:  0 ['after', 'many', 'hours', 'working', 'in', 'medical', 'isolation', 'ward', 'a', 'picture', 'of', 'an', 'egyptian', 'hero', 'from', 'egyptian', 'white', 'army', '.']\n",
            "pred:  1 label:  1 ['a', 'study', 'from', 'annals', 'of', 'internal', 'medicine', 'estimates', 'that', 'probability', 'of', 'pcr', 'tests', 'resulting', 'in', 'varies', 'over', 'course', 'of', 'illness', 'highest', 'risk', 'being', 'before', 'symptoms', 'appear', 'and', 'lowest', 'three', 'days', 'after', 'symptoms', 'appear', '.', '.']\n",
            "pred:  1 label:  1 ['case', 'means', 'our', 'total', 'number', 'of', 'confirmed', 'cases', 'of', 'is', '1154', 'which', 'is', 'number', 'we', 'report', 'to', 'world', 'health', 'organization', '.', 'our', 'combined', 'total', 'of', 'confirmed', 'and', 'probable', 'cases', 'is', '1504', '.']\n",
            "pred:  1 label:  1 ['our', '4', 'pm', 'update', 'is', 'published', '.', 'us', 'has', 'now', 'completed', 'tests', 'on', 'at', 'least', 'people', 'up', 'from', 'total', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'and', 'not', 'all', 'states', 'report', 'all', 'negative', 'tests', '.', 'for', 'details', '.']\n",
            "pred:  0 label:  0 ['dr', '.', 'vele', 'markovski', 'that', 'ventilators', 'and', 'quarantines', 'are', 'doing', 'great', 'harm', 'in', 'treating', 'coronavirus', 'and', 'that', 'contamination', 'level', 'of', 'is', 'extremely', 'low']\n",
            "pred:  1 label:  1 ['5/10', 'boris', 'johnson', 'will', 'chair', 'an', 'emergency', 'cobra', 'meeting', 'tomorrow', 'ahead', 'of', 'a', 'statement', 'on', 'next', 'steps', 'in', 'tackling', 'pandemic', '.', 'click', 'below', 'to', 'find', 'out', 'what', 'pm', 'could', 'announce', '.']\n",
            "pred:  1 label:  1 ['boris', 'johnson', 'has', 'said', 'members', 'of', 'public', 'should', 'not', 'report', 'neighbours', 'who', 'breach', 'of', 'unless', 'they', 'are', 'having', 'parties', 'with', 'tubs', 'and', 'so', '.']\n",
            "pred:  0 label:  0 ['vaccines', 'killed', '50', 'million', 'people', 'during', '1918', 'spanish', 'flu', 'pandemic']\n",
            "pred:  0 label:  0 ['after', 'months', 'of', 'decreased', 'pollution', 'with', 'humans', 'in', 'lockdown', ',', 'gorilla', 'populations', 'are', 'starting', 'to', 'develop', 'rudimentary', 'nuclear', 'weapons']\n",
            "pred:  1 label:  1 ['last', 'line', 'of', 'defence', 'is', 'full', 'national', 'action', '.', '”', 'health', 'secretary', 'says', 'he', 'not', 'want', 'to', 'a', 'second', 'national', 'lockdown', 'but', 'government', 'will', 'do', 'what', 'is', 'necessary', 'to', 'people', '.', '.', '.']\n",
            "pred:  1 label:  1 ['good', 'news', 'remains', 'falling', 'number', 'of', 'reported', 'deaths', 'as', '7-day', 'average', 'continues', 'to', '.']\n",
            "pred:  0 label:  0 ['in', 'philippines', 'bagong', 'makabayan', 'secretary', 'general', 'renato', 'reyes', 'was', 'spotted', 'panic', 'buying', 'in', 'an', 'branch', '.']\n",
            "pred:  0 label:  0 ['coffee', 'can', 'cure', '.']\n",
            "pred:  1 label:  1 ['a', 'small', 'fraction', 'of', 'deaths', 'in', 'long-term', 'care', 'facilities', 'are', 'staff', 'members', '.', 'nonetheless', 'researchers', 'estimate', 'that', 'by', 'end', 'will', 'make', 'working', 'in', 'an', 'ltc', 'facility', 'most', 'dangerous', 'job', 'in', 'america', 'in', '2020', '.']\n",
            "pred:  0 label:  0 ['elizabeth', 'ii', 'has', 'been', 'diagnosed', 'with', 'coronavirus', '.', '_']\n",
            "pred:  0 label:  0 ['is', 'disrupting', 'pensions', '.']\n",
            "pred:  0 label:  0 ['bill', 'gates', 'said', 'that', 'catholic', 'religion', 'may', 'have', 'to', 'be', 'suspended', 'forever', 'because', 'of', 'pandemic', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'fish', 'tank', 'additive', 'may', 'treat', 'coronavirus', '.', '?', '?']\n",
            "pred:  1 label:  1 ['cases', 'in', 'india', 'on', 'september', '7', '.', '.', 'active', 'cases', '.', 'deaths', 'total', 'confirmed', 'cases', '=', '.']\n",
            "pred:  1 label:  1 ['rise', 'in', 'deaths', 'is', 'largely', 'driven', 'by', 'southern', 'states', 'which', 'reported', '962', 'deaths', 'today', '.', '.']\n",
            "pred:  1 label:  1 ['there', 'has', 'been', 'a', 'steep', 'exponential', 'rise', 'in', 'from', '50000', 'in', 'may', 'to', '30', 'lakh', 'in', 'sept', '.', 'daily', 'number', 'of', 'recovered', 'patients', 'has', 'crossed', '70000', '.', 'more', 'than', '3/4', 'of', 'total', 'cases', 'have', 'recovered', '.', '.']\n",
            "pred:  1 label:  1 ['georgia', 'also', 'reported', 'a', 'record', 'high', 'in', 'cases', 'at', 'almost', '2000', '.', '7-day', 'average', 'for', 'cases', 'has', 'nearly', 'tripled', 'since', 'late', 'may', '.', 'hospitalizations', 'have', 'also', 'gone', 'right', 'back', 'up', 'after', 'falling', 'from', 'may', 'into', 'june', '.', '.']\n",
            "pred:  0 label:  0 ['proud', 'of', 'team', 'busy', 'in', 'packing', 'stuff', 'for', 'our', '19', 'family', 'support', '.', 'ensuring', 'physical', 'distancing', 'as', 'well', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'total', 'confirmed', 'cases', 'cases', 'samples', 'tested']\n",
            "pred:  0 label:  0 ['trump', 'hails', 'talks', 'with', '.']\n",
            "pred:  0 label:  0 ['garlic', 'water', 'can', 'cure', 'novel', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['trial', 'of', 'homegrown', 'vaccine', 'soon', 'in', 'these', 'two', 'cities', ',', 'know', 'here', 'in', 'detail', '-', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'governments', 'across', 'country', 'are', 'adding', 'thousands', 'of', 'jobs', 'in', 'field', 'of', 'contact', 'tracing', 'to', 'fight', '.']\n",
            "pred:  1 label:  1 ['239', 'new', 'cases', 'of', ';', 'confirmed', 'discharged', '1023', 'deaths', '.']\n",
            "pred:  1 label:  1 ['this', 'is', 'a', 'critical', 'moment', 'for', 'countries', 'and', 'we', 'ask', 'leaders', 'to', 'put', 'targeted', 'measures', 'in', 'place', 'that', 'we', 'know', 'can', 'suppress', 'spread', 'and', 'ensure', 'that', 'health', 'systems', 'and', 'workers', 'are', 'protected', '.']\n",
            "pred:  0 label:  0 ['really', ',', 'meat', 'packing', 'plants', 'are', 'asks', 'smug', 'ghost', 'of', 'upton', 'sinclair', '.']\n",
            "pred:  1 label:  1 ['ai', 'spots', 'lung', 'damage', 'with', 'accuracy', '.', '.']\n",
            "pred:  0 label:  0 ['no', 'osha', 'federal', 'workplace', 'health', 'and', 'safety', 'agency', 'does', 'not', 'say', 'that', 'masks', 'work', 'against', '.', '.']\n",
            "pred:  1 label:  1 ['planning', 'on', 'visiting', 'family', 'or', 'friends', 'over', 'weekend', '?', 'protect', 'them', 'and', 'yourself', 'from', 'wear', 'a', 'mask', 'stay', 'at', 'least', '6', 'feet', 'away', 'from', 'people', 'you', 'live', 'with', 'and', 'wash', 'your', 'hands', 'often', '.', '.', '.', '.']\n",
            "pred:  0 label:  0 ['no', 'evidence', 'that', 'an', 'american', 'vaccine', 'killed', '5', 'ukrainians', '.', 'claim', 'comes', 'from', 'a', 'site', 'known', 'for', 'disinformation', '.', '.']\n",
            "pred:  1 label:  1 ['last', 'data', 'we', 'are', 'not', 'reporting', 'pending', 'number', 'because', 'it', 'has', 'been', 'published', 'irregularly', '.', 'but', 'it', 'is', 'quite', 'worth', 'considering', 'when', 'looking', 'at', 'data', '.', ')']\n",
            "pred:  0 label:  0 ['president', 'trump', 'misconstrued', 'data', 'on', 'coronavirus', 'deaths', '.', 'as', 'of', 'sept', '.', '3', 'cdc', 'data', 'show', 'americans', 'had', 'died', 'due', 'to', 'and', 'some', 'estimates', 'put', 'death', 'toll', 'higher', '.', '.']\n",
            "pred:  1 label:  1 ['3', 'vaccines', 'are', 'in', 'progress', 'and', '3', 'more', 'vaccines', 'are', 'in', 'pre-clinical', 'dg', '.']\n",
            "pred:  1 label:  1 ['all', 'those', 'in', 'hotel', 'still', 'were', 'tested', 'over', 'last', 'day', 'and', 'a', 'half', 'and', 'those', 'tests', 'are', 'coming', 'through', 'mostly', 'today', '.', 'no', 'one', 'is', 'to', 'leave', 'managed', 'isolation', 'facilities', 'unless', 'they', 'have', 'had', 'a', 'negative', 'test', '.', 'day', '3', 'and', 'day', '12', 'testing', 'is', 'in', 'full', 'swing', '.']\n",
            "pred:  0 label:  0 ['does', 'reduce', 'your', 'chances', 'of', 'getting', 'coronavirus', '?']\n",
            "pred:  0 label:  0 ['department', 'of', 'health', 'reports', 'fake', 'cigarettes', 'that', 'spread', 'coronavirus', 'have', 'reached', 'philippines', '.']\n",
            "pred:  1 label:  1 ['less', 'than', 'half', 'states', 'are', 'reporting', 'hospitalization', 'numbers', 'in', 'any', 'way', '.', 'washington', 'is', 'among', 'states', 'missing', 'and', 'presumably', 'has', 'second', 'largest', 'number', 'of', 'hospitalized', 'people', 'of', 'severity', 'and', 'duration', 'of', 'its', '.']\n",
            "pred:  0 label:  0 ['a', 'banner', 'with', 'a', 'swastika', 'and', 'is', 'from', 'a', 'michigan', 'coronavirus', 'protest', '.']\n",
            "pred:  1 label:  1 ['5', 'new', 'confirmed', 'cases', 'in', 'community', 'are', 'all', 'clearly', 'epidemiologically', 'linked', 'to', 'auckland', 'cluster', '.', '3', 'cases', 'are', 'linked', 'to', 'mt', 'roskill', 'church', 'other', '2', 'are', 'household', 'contacts', 'of', 'two', 'separate', 'previously', 'reported', 'cases', '.']\n",
            "pred:  0 label:  0 ['number', 'of', 'deaths', 'in', 'brazil', 'plummeted', 'after', 'minister', 'of', 'justice', 'sergio', 'moro', 'ordered', 'federal', 'police', 'to', 'investigate', 'whether', 'numbers', 'were', 'being', 'fabricated', '.']\n",
            "pred:  1 label:  1 ['grandparents', 'and', 'others', 'who', 'provide', 'informal', 'childcare', 'will', 'be', 'exempt', 'from', 'coronavirus', 'rules', 'in', 'local', 'lockdown', 'areas', 'in', 'england', '.']\n",
            "pred:  0 label:  0 ['russia', 'reports', 'huge', 'spike', 'in', 'coronavirus', 'deaths', 'among', 'journalists', '.']\n",
            "pred:  0 label:  0 ['medical', 'genius', 'donald', 'trump', 'discovers', 'that', 'infection', 'numbers', 'go', 'down', 'when', 'you', 'test', 'for', 'it', '.', '.']\n",
            "pred:  0 label:  0 ['says', 'vatican', 'confirmed', 'that', 'pope', 'francis', 'and', 'two', 'aides', 'tested', 'positive', 'for', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['.', 'has', 'improved', 'reporting', 'on', 'cause', 'of', 'death', 'which', 'is', 'an', 'important', 'first', 'step', 'towards', 'real-time', 'data', '.', 'highlights', 'need', 'for', 'accurate', 'timely', 'data', 'more', 'than', 'ever', '.', '.']\n",
            "pred:  1 label:  1 ['intensive', 'care', 'units', 'are', 'either', 'full', 'or', 'about', 'to', 'reach', 'saturation', '.', 'more', 'bed', 'capacity', 'is', 'desperately', 'needed', 'but', 'staffing', 'is', 'an', 'issue', 'witnesses', 'impact', 'a', 'second', 'wave', 'of', 'is', 'having', 'is', 'having', 'in', 'marseille', '.']\n",
            "pred:  0 label:  0 ['video', 'shows', 'muslim', 'women', 'spitting', 'in', 'plastic', 'bags', 'and', 'throwing', 'them', 'into', 'houses', 'to', 'spread', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['a', 'case', 'of', 'novel', 'coronavirus', 'was', 'registered', 'in', 'a', 'hospital', 'in', 'maranhão', 'brazil', '.']\n",
            "pred:  1 label:  1 ['having', 'a', 'difficult', 'time', 'coping', 'because', 'of', '?', 'feeling', 'lonely', 'anxious', 'or', 'dealing', 'with', 'grief', 'and', 'loss', '?', 'help', 'is', 'available', 'through', 'national', 'prevention', 'lifeline', '.', 'learn', '.', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'photo', 'shows', 'three', 'men', 'on', 'crosses', 'who', '?', '?', '?', 'feel', 'like', 'they', '?', '?', '?', 're', 'being', 'crucified', 'by', 'being', 'required', 'to', 'stay', 'at', 'home', '.', '?', '?']\n",
            "pred:  0 label:  0 ['meghan', 'markle', 'is', 'donating', 'one', 'of', 'her', 'royal', 'crowns', 'to', 'cause', '.']\n",
            "pred:  0 label:  0 ['riding', 'in', 'time', 'of', 'how', 'single', 'people', 'are', 'handling', 'things', '.']\n",
            "pred:  1 label:  1 ['rt', 'update', 'gap', 'between', 'recovered', ';', 'active', 'cases', 'has', 'crossed', '29', 'lakh', 'today', 'cases', 'are', 'nearly', '4']\n",
            "pred:  0 label:  0 ['could', 'miley', 'vagina', 'hold', 'clue', 'to', 'vaccine', '?', '.']\n",
            "pred:  0 label:  0 ['has', 'no', 'deaths', 'from', ';', 'tea', 'made', 'of', 'lemon', 'and', 'bicarbonate', 'can', 'cure', 'coronavirus', '.', '_']\n",
            "pred:  1 label:  1 ['are', 'very', 'concerned', 'about', 'exponential', 'rise', 'in', 'number', 'of', 'cases', 'across', 'north', 'east', '.', '”', 'newcastle', 'council', 'leader', 'explains', 'need', 'for', 'new', 'restrictions', 'in', 'north', 'east', 'of', 'england', '.', '.', '.']\n",
            "pred:  0 label:  0 ['decision', 'of', 'lockdown', 'was', 'timely', ',', 'curve', 'has', 'begun', 'to', 'flatten', 'nation', 'has', 'shown', 'that', 'lockdown', 'has', 'been', 'effective', ',', ',', 'containing', 'infection', 'and', 'slowing', 'down', 'doubling', 'rate', 'we', 'need', 'to', 'build', 'on', 'these', 'gains']\n",
            "pred:  0 label:  0 ['prolonged', 'use', 'of', 'mask', 'causes', 'hypoxia', '.']\n",
            "pred:  1 label:  1 ['there', 'are', '54', 'people', 'linked', 'to', 'community', 'cluster', 'who', 'remain', 'in', 'auckland', 'quarantine', 'facility', 'which', 'includes', '23', 'people', 'who', 'have', 'tested', 'positive', 'for', 'and', 'their', 'household', 'contacts', '.']\n",
            "pred:  1 label:  1 ['as', 'of', 'june', '15', 'forecasts', 'suggest', 'number', 'of', 'deaths', 'will', 'continue', 'to', 'slow', 'nationally', 'with', 'to', 'deaths', 'by', 'july', '11', '.', 'however', '9', 'states', 'are', 'likely', 'to', 'report', 'more', 'deaths', 'in', 'next', '4', 'weeks', 'than', 'previous', '4', '.', '.']\n",
            "pred:  1 label:  1 ['we', 'are', 'hoping', 'for', 'full', 'recovery', '.', 'there', 'are', 'now', '160', 'people', 'linked', 'to', 'cluster', 'they', 'have', 'moved', 'into', 'managed', 'quarantine', 'facility', 'in', 'auckland', 'and', 'that', 'includes', '89', 'people', 'who', 'have', 'returned', 'positive', 'tests', 'as', 'well', 'as', 'household', 'contacts', '.']\n",
            "pred:  0 label:  0 ['video', 'shows', 'a', 'new', 'hospital', 'for', 'coronavirus', 'patients', 'in', 'china', '.', '_']\n",
            "pred:  1 label:  1 ['daily', 'testing', 'capacity', 'has', 'crossed', '11', '.', '70', 'lakh', ';', 'cumulative', 'tests', 'are', 'nearly', '5', 'crore', 'as', 'on', 'date', '.', 'tests', 'were', 'conducted', 'in', 'last', '24', 'hours', '.', '.']\n",
            "pred:  0 label:  0 ['first', 'people', 'to', 'receive', 'russian', 'coronavirus', 'vaccine', 'now', 'able', 'to', 'run', '100m', 'in', 'under', 'ten', 'seconds', '.', '.']\n",
            "pred:  1 label:  1 ['kia', 'ora', 'there', 'are', 'currently', 'no', 'active', 'cases', 'in', 'wellington', '.']\n",
            "pred:  1 label:  1 ['to', 'support', 'countries', 'in', 'every', 'situation', 'and', 'who', 'yesterday', 'published', 'updated', 'guidance', 'on', 'school-related', 'public', 'health', 'measures', 'in', 'context', 'of', '.']\n",
            "pred:  0 label:  0 ['quarantine', 'style', '.', '.', '.', 'shot', 'dead', '.', '.']\n",
            "pred:  0 label:  0 ['looters', 'maintain', 'social', 'distancing', 'while', 'creating', 'anarchy', '.']\n",
            "pred:  1 label:  1 ['rt', 'wanted', 'to', 'talk', 'a', 'little', 'about', 'data', 'gathering', 'and', 'checking', 'methodology', 'which', 'building', 'as']\n",
            "pred:  1 label:  1 ['from', 'endless', 'warnings', 'about', 'world', 'being', 'all', 'countries', 'need', 'to', 'dig', 'in', 'together', 'and', 'invest', 'to', 'ensure', 'a', 'pandemic', 'of', 'this', 'magnitude', 'and', 'severity', 'never', 'happens', 'again', '.']\n",
            "pred:  1 label:  1 ['there', 'are', '427', 'people', 'who', 'we', 'have', 'repeatedly', 'tried', 'to', 'make', 'contact', 'with', 'including', 'via', 'text', 'and', 'via', 'phone', 'calls', '.', 'again', 'a', 'reminder', 'to', 'anyone', 'who', 'was', 'in', 'a', 'managed', 'isolation', 'facility', 'between', 'june', '9-16', 'who', 'has', 'not', 'yet', 'spoken', 'with', 'healthline', 'to', 'call', 'dedicated', 'team', 'on', '09', '302', '.']\n",
            "pred:  1 label:  1 ['there', 'are', 'no', 'additional', 'deaths', 'to', 'report', '.', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'is', '.', 'today', 'is', 'world', 'family', 'doctor', 'day', '.', 'a', 'big', 'thanks', 'to', '5500', 'gps', 'around', 'new', 'zealand', 'for', 'work', 'they', 'do', 'every', 'day', '.']\n",
            "pred:  1 label:  1 ['here', 'is', 'map', 'of', 'new', 'cases', '.', '.']\n",
            "pred:  0 label:  0 ['40', 'flight', 'crash', 'victims', 'test', 'positive', 'for', '19']\n",
            "pred:  1 label:  1 ['rt', 'if', 'you', 'are', 'struggling', 'with', 'stress', 'or', 'mental', 'health', 'concerns', 'due', 'to', 'or', 'after', 'a', 'disaster', 'there', 'is', 'help', 'available', '.']\n",
            "pred:  1 label:  1 ['acc', 'to', 'may', 'persist', 'on', 'surfaces', 'for', 'few', 'hrs', 'to', 'up', 'to', 'several', 'days', '.', 'this', 'duration', 'may', 'vary', 'under', 'different', 'conditions', 'type', 'of', 'surface', 'temperature', 'humidity', '.', '.']\n",
            "pred:  0 label:  0 ['immunologist', 'and', 'nobel', 'winner', 'tasuku', 'honjo', 'said', 'new', 'coronavirus', 'natural', 'and', 'was', 'created', 'by', 'chinese', '.']\n",
            "pred:  0 label:  0 ['videos', 'showing', 'black', 'people', 'being', 'attacked', 'by', 'asian', 'people', 'have', 'been', 'shared', 'thousands', 'of', 'times', 'online', 'in', 'recent', 'weeks', '.', 'although', 'africans', 'living', 'in', 'china', 'have', 'reported', 'discrimination', 'linked', 'to', 'coronavirus', 'pandemic']\n",
            "pred:  0 label:  0 ['is', 'decision', 'of', 'president', ',', '\"', 'not', 'governors', ',', 'to', 'up', 'states', '.', '\"']\n",
            "pred:  1 label:  1 ['100', 'new', 'cases', 'of', ';', 'confirmed', 'discharged', '1057', 'deaths', '.']\n",
            "pred:  0 label:  0 ['u', '.', 's', '.', 'administration', 'under', 'president', 'obama', 'gave', 'wuhan', 'institute', 'of', 'virology', 'us$', '3', '.', '7', 'million']\n",
            "pred:  1 label:  1 ['and', 'through', 'access', 'to', 'tools', 'accelerator', 'and', 'global', 'vaccines', 'facility', 'working', 'to', 'ensure', 'that', 'if', 'and', 'when', 'a', 'vaccine', 'is', 'proven', 'to', 'be', 'safe', 'and', 'effective', 'it', 'will', 'be', 'accessible', 'equitably', 'for', 'all', 'countries', '.']\n",
            "pred:  1 label:  1 ['as', 'of', 'today', 'in', 'is', '81', '.', '.', '.']\n",
            "pred:  0 label:  0 ['videos', 'of', 'empty', 'hospitals', 'show', 'is', 'fake', '.']\n",
            "pred:  1 label:  1 ['we', 'still', 'have', '16', 'significant', 'clusters', 'four', 'of', 'which', 'are', 'now', 'closed', '.', 'there', 'are', 'no', 'additional', 'deaths', 'to', 'report', '.', 'yesterday', 'our', 'labs', 'processed', 'tests', 'total', 'number', 'of', 'tests', 'to', 'date', 'is', '.', 'there', 'will', 'be', 'a', 'further', 'update', 'from', 'government', 'at', '4', 'pm', '.']\n",
            "pred:  1 label:  1 ['nearly', 'of', 'total', 'cases', 'are', 'being', 'contributed', 'by', 'five', 'states', 'viz', '.', 'maharashtra', 'andhra', 'pradesh', 'tamil', 'nadu', 'karnataka', 'and', 'uttar', 'pradesh', '.', '.']\n",
            "pred:  1 label:  1 ['due', 'to', 'recent', 'low', 'testing', 'numbers', 'likely', 'related', 'to', 'holiday', 'weekend', '7-day', 'average', 'tests', 'fell', 'to', 'lowest', 'since', 'mid-july', '.', '.']\n",
            "pred:  1 label:  1 ['multisystem', 'inflammatory', 'syndrome', 'in', 'children', 'is', 'a', 'rare', 'but', 'serious', 'condition', 'associated', 'with', '.', 'for', 'more', 'information', 'about', 'and', 'when', 'to', 'seek', 'emergency', 'care', 'for', 'your', 'child', 'visit', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'screen', 'capture', 'of', 'cover', 'of', 'new', 'york', 'times', 'accompanied', 'by', 'a', 'text', 'in', 'which', 'it', 'is', 'assured', 'that', 'newspaper', 'would', 'have', 'branded', 'government', 'of', 'spain', 'as', '.']\n",
            "pred:  1 label:  1 ['while', 'grateful', 'for', 'funds', 'already', 'committed', 'towards', 'facility', 'more', 'is', 'urgently', 'needed', 'to', 'continue', 'to', 'move', 'portfolio', 'forward', '.', '.']\n",
            "pred:  0 label:  0 ['as', 'india', 'detected', 'more', 'cases', 'in', 'march-april', '2020', 'fake', 'misleading', 'claims', 'around', 'rose', '.', 'most', 'common', 'were', 'communal', 'rumours', 'followed', 'by', 'false', 'per', 'analysis', '.', '.']\n",
            "pred:  1 label:  1 ['while', 'most', 'attention', 'has', 'been', 'focused', 'on', 'big', 'outbreaks', 'across', 'southeast', 'and', 'in', 'arizona', 'there', 'are', 'several', 'states', 'outside', 'region', 'that', 'look', 'to', 'be', 'on', 'verge', 'of', 'seeing', 'much', 'higher', 'levels', 'of', 'transmission', '.', '.']\n",
            "pred:  0 label:  0 ['list', 'shows', 'medicines', 'issued', 'by', 'a', 'doctor', 'in', 'gangaram', 'hospital', 'to', 'alleviate', 'coronavirus', 'symptoms', '.']\n",
            "pred:  1 label:  1 ['i', 'want', 'to', 'thank', 'and', 'for', 'announcing', 'that', 'they', 'will', 'be', 'joining', '.', 'my', 'thanks', 'to', 'other', 'countries', 'who', 'are', 'making', 'commitments', 'ahead', 'of', 'deadline', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '17', '.', '8', 'million', 'tests', 'up', 'from', 'yesterday', '.', '7-day', 'average', 'is', 'now', 'over', '400k', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'for', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['persistence', 'of', 'long-term', 'symptoms', 'in', 'some', 'patients', 'has', 'opened', 'up', 'a', 'new', 'line', 'of', 'research', 'into', 'mechanisms', 'underlying', 'as', 'well', 'as', 'other', 'chronic', 'illnesses', '.', '.']\n",
            "pred:  0 label:  0 ['donald', 'trump', 'is', 'not', 'responsible', 'for', ',', 'but', 'he', 'does', 'bear', 'full', 'responsibility', 'for', 'his', 'slow', ',', 'uncoordinated', ',', 'bungled', 'response', '.', '.']\n",
            "pred:  1 label:  1 ['your', 'look', 'at', 'number', 'of', 'new', 'cases', 'per', 'million', '.', 'louisiana', 'has', 'stopped', 'reporting', 'on', 'saturdays', 'so', 'two', 'days', 'worth', 'of', 'cases', '.', '.']\n",
            "pred:  0 label:  0 ['tablighi', 'jamaat', 'had', 'asked', 'for', 'non-vegetarian', 'food', 'and', 'defecated', 'in', 'open', 'at', 'a', 'quarantine', 'facility', 'located', 'in', 'saharanpur', ',', 'uttar', 'pradesh', '.']\n",
            "pred:  0 label:  0 ['nokia', 'distributing', 'free', 'phones', 'to', 'students', 'amidst', 'coronavirus', 'pandemic']\n",
            "pred:  1 label:  1 ['newly', 'investigational', 'drugs', 'and', 'therapies', '.', '.', '.']\n",
            "pred:  1 label:  1 ['yesterday', 'hhs', 'sec', '.', 'azar', 'announced', 'has', 'completed', 'its', 'review', 'of', 'ind', 'application', 'for', 'our', 'potential', 'vaccine', 'for', ';', 'allowed', 'it', 'to', 'proceed', 'to', 'clinic', '.', 'on', 'feb', '.', '24', '1st', 'clinical', 'batch', 'of', 'was', 'shipped', 'to', 'nih', 'for', 'use', 'in', 'phase', '1', 'study', '.']\n",
            "pred:  1 label:  1 ['establishes', 'its', 'global', 'position', 'with', 'export', 'of', '23', 'lakh', 'in', 'one', 'month', '.', 'more', 'than', '1', '.', '28', 'cr', 'distributed', 'to', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'popular', 'florida', 'beach', 'has', 'been', 'evacuated', 'after', 'jellyfish', 'test', 'positive', 'for', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['everyone', 'arriving', 'in', 'zimbabwe', 'will', 'now', 'have', 'to', 'pay', 'for', 'a', 'pcr', 'test']\n",
            "pred:  1 label:  1 ['also', 'shipping', 'millions', 'of', 'items', 'of', 'protective', 'gear', 'around', 'world', 'and', 'ensuring', 'health', 'facilities', 'are', 'properly']\n",
            "pred:  0 label:  0 ['coronavirus', 'up', 'on', 'adding', 'that', 'it', 'is', 'very', 'unforeseen', 'thing', '.', '”']\n",
            "pred:  0 label:  0 ['president', 'said', 'that', 'country', 'has', '10000', 'tests', '.']\n",
            "pred:  1 label:  1 ['newest', 'cdc', 'health', 'alert', 'network', 'update', 'recommends', 'all', 'people', 'defer', 'any', 'travel', 'on', 'cruise', 'ships', 'including', 'river', 'cruises', 'worldwide', 'because', 'of', 'increased', 'risk', 'of', 'spread', 'onboard', 'ships', '.', 'learn', '.', '.']\n",
            "pred:  0 label:  0 ['did', 'you', 'see', 'those', 'clive', 'palmer', 'ads', 'as', 'a', 'treatment', 'for', 'in', 'newscorp', 'papers', 'recently', '?', 'runs', 'rule', 'over', 'claims', 'in', 'ad', 'for', 'on', '.']\n",
            "pred:  1 label:  1 ['india', 'reports', 'highest', 'ever', 'single', 'day', 'recoveries', 'of', 'patients', '.', '.']\n",
            "pred:  0 label:  0 ['donald', 'trump', 'forgot', 'about', ',', 'but', 'forget', 'about', 'us', '.', 'president', 'wish', 'away', 'virus', 'in', 'april', ',', 'he', 'tweet', 'it', 'away', 'in', 'may', ',', 'and', 'he', 'ignore', 'it', 'away', 'in', 'june', '.']\n",
            "pred:  0 label:  0 ['little', 'boy', 'died', 'of', 'in', 'three', 'different', 'countries', '.', 'still', 'believe', 'media', 'is', '?', '”']\n",
            "pred:  0 label:  0 ['author', 'or', 'authors', 'of', 'leaflet', 'remain', 'anonymous', '.', 'if', 'you', 'have', 'any', 'information', 'on', 'source', 'or', 'have', 'seen', 'any', 'other', 'similar', 'leaflets', 'in', 'your', 'mailbox', 'tweet', 'them', '@', 'us', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'of', 'a', '14-year-old', 'indian', 'boy', 'named', 'anand', 'who', 'was', 'alleged', 'to', 'have', 'predicted', 'coronavirus', 'in', '2019', '.']\n",
            "pred:  0 label:  0 ['ugandan', 'president', 'announces', 'postponement', 'of', 'elections', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'now', ',', 'they', '?', '?', '?', 're', 'doing', 'tests', 'on', 'airlines', '?', '?', '\"', 'very', 'strong', 'tests', '?', '?', '\"', 'for', 'getting', 'on', ',', 'getting', 'off', '.', 'they', '?', '?', '?', 're', 'doing', 'tests', 'on', 'trains', '?', '?', '\"', 'getting', 'on', ',', 'getting', 'off', '.', '?', '?']\n",
            "pred:  0 label:  0 ['bill', 'gates', 'said', 'of', 'people', 'will', 'die', 'with']\n",
            "pred:  0 label:  0 ['world', 'health', 'organization', 'has', 'advised', 'people', 'against', 'eating', 'bakery', 'items', 'amid', 'outbreak', '.']\n",
            "pred:  1 label:  1 ['with', 'this', 'steep', 'exponential', 'rise', 'in', 'daily', 'testing', 'cumulative', 'tests', 'are', 'nearing', '4', '.', '7', 'crore', '.', 'cumulative', 'tests', 'as', 'on', 'date', 'have', 'reached', 'details', ':', '.']\n",
            "pred:  1 label:  1 ['6', 'potential', 'treatments', 'identified', 'by', 'ai', '.', '.']\n",
            "pred:  0 label:  0 ['atheist', 'italian', 'doctor', 'julian', 'urban', 'and', 'his', 'colleagues', 'have', 'converted', 'to', 'christianity', 'in', 'wake', 'of', 'crisis', '.']\n",
            "pred:  1 label:  1 ['total', 'recoveries', 'have', 'crossed', '30', 'lakh', 'today', '.', 'india', 'has', 'continued', 'its', 'trajectory', 'of', 'posting', 'more', 'than', '60000', 'recoveries', 'for', '8th', 'consecutive', 'day', '.', 'recovery', 'rate', 'reaches', 'to', '77', '.', '.', '.']\n",
            "pred:  0 label:  0 ['despite', 'pandemic', 'total', 'number', 'of', 'deaths', 'in', 'brazil', 'decreased', 'rather', 'than', 'increased', 'when', 'we', 'compare', 'months', 'of', 'april', '2019', 'and', 'april', '2020', '.', 'this', 'statement', 'is', 'backed', 'by', 'national', 'civil', 'registry', 'data', '.']\n",
            "pred:  1 label:  1 ['cases', 'in', 'bolton', 'are', 'continuing', 'to', 'increase', 'with', 'equivalent', 'of', '192', 'new', 'cases', 'per', '100000', 'people', 'recorded', 'in', 'week', 'to', '10', 'september', '.', 'latest', 'on', 'this', 'breaking', '.', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'up', 'on', 'us', ',', '?', '?', 'adding', 'that', 'it', 'is', '?', '?', '?', 'a', 'very', 'unforeseen', 'thing', '.', '?', '?']\n",
            "pred:  0 label:  0 ['data', 'scientist', 'who', 'designed', 'dashboard', '—', 'a', 'mobile', 'friendly', ',', 'has', 'been', 'removed', 'from', 'her', 'position', 'because', 'she', 'refused', 'to', 'censor', 'data', 'and', 'manipulate', 'numbers', '.', '.', '.']\n",
            "pred:  0 label:  0 ['bill', 'gates', 'personally', 'will', 'profit', 'from', 'a', 'vaccine', 'and', 'he', 'owns', 'a', 'company', 'that', 'plans', 'to', 'implant', 'microchips', 'in', 'everyone', '.']\n",
            "pred:  1 label:  1 ['upsurge', 'in', 'testing', 'india', 'crosses', 'a', 'new', 'peak', 'of', '4', 'crore', 'tests', 'than', '9', 'lakh', 'tests', 'conducted', 'for', 'third', 'successive', 'day', '.', '.']\n",
            "pred:  1 label:  1 ['going', 'to', 'see', 'more', 'transmission', 'going', 'forward', '.', '.', 'via', '.']\n",
            "pred:  1 label:  1 ['in', 'mainland', 'early', 'implementation', ';', 'timely', 'adjustment', 'of', 'control', 'measures', 'important', 'to', 'contain', 'transmission', 'data', 'made', 'publicly', 'available', 'as', 'additional', 'source', 'for', 'research', ';', 'policy', 'planning', '.', '.']\n",
            "pred:  1 label:  1 ['testing', 'in', 'india', 'has', 'steeply', 'increased', 'to', 'more', 'than', '9', 'lakh', 'tests', 'per', 'day', '.', 'this', 'is', 'bolstered', 'with', 'all', 'testing', 'more', 'than', '140', 'population', 'as', 'advised', 'by', '.', '.']\n",
            "pred:  1 label:  1 ['they', 'were', 'tested', 'as', 'part', 'of', 'routine', 'requirements', 'for', 'their', 'safe', 'return', 'to', 'work', 'and', 'results', 'came', 'back', 'positive', '.', 'they', 'remain', 'in', 'isolation', '.', 'third', 'case', 'is', 'a', 'person', 'who', 'has', 'travelled', 'back', 'from', 'overseas', 'so', 'is', 'an', 'imported', 'case', '.']\n",
            "pred:  0 label:  0 ['ratan', 'view', 'on', 'how', 'indian', 'economy', 'will', 'bounce', 'back', 'if', 'there', 'is', 'an', 'economic', 'downfall', 'due', 'to', 'current', 'situation', '.']\n",
            "pred:  1 label:  1 ['also', 'mitigating', 'impact', 'addressing', 'psychosocial', 'needs', 'and', 'facilitating', 'rapid', 'restoration', 'of', 'damaged', 'health', 'facilities', '.', 'we', 'have', 'issued', 'an', 'appeal', 'for', 'million', 'and', 'ask', 'for', 'your', 'solidarity', 'and', 'support', 'to', 'lebanese']\n",
            "pred:  1 label:  1 ['rt', 'update', '1132', 'case', 'fatalities', 'have', 'been', 'reported', 'in', 'past', '24', 'hrs', '.', 'with', '474', 'new', 'cases', 'of', 'mortality', 'in']\n",
            "pred:  1 label:  1 ['icmr', 'testing', 'crossed', '3', 'lakh', '.', 'for', 'more', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['before', 'this', 'change', 'we', 'indicated', 'that', 'mass', '.', 'was', 'at', 'an', 'elevated', 'risk', '.', 'after', 'shifting', 'our', 'scale', 'state', 'is', 'now', 'meaning', 'there', 'is', 'risk', 'of', 'second', 'spike', 'or', 'major', 'gap', 'in', 'at', 'least', 'one', 'of', 'metrics', 'and', 'can', 'quickly', 'devolve', 'to', 'an', 'imminent', 'outbreak', 'barring', 'intervention', '.']\n",
            "pred:  1 label:  1 ['florida', 'reported', 'a', 'record', '277', 'deaths', 'today', '.', 'given', 'reporting', 'problems', 'there', 'over', 'last', 'week', 'there', 'may', 'be', 'more', 'deaths', 'from', 'further', 'in', 'past', 'than', 'on', 'a', 'standard', 'day', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'update', 'total', 'number', 'of', 'recoveries', 'has', 'touched', 'taking', 'recovery', 'rate', 'to', '77', '.']\n",
            "pred:  0 label:  0 ['posts', 'saying', 'that', 'for', 'several', 'centuries', 'there', 'have', 'been', 'pandemics', 'in', '.']\n",
            "pred:  1 label:  1 ['one', 'case', 'previously', 'reported', 'in', 'kano', 'has', 'been', 'transferred', 'to', 'jigawa', 'state', '.', 'therefore', 'total', 'number', 'of', 'confirmed', 'cases', 'in', 'kano', 'is', '36', 'as', 'at', '19th', 'of', 'april', '2020']\n",
            "pred:  0 label:  0 ['children', 'will', 'be', 'separated', 'from', 'parents', 'for', 'coronavirus', 'treatment', '.']\n",
            "pred:  1 label:  1 ['people', 'can', 'spread', 'to', 'pets', '.', 'protect', 'your', 'dog', 'by', 'limiting', 'their', 'contact', 'with', 'people', 'outside', 'your', 'household', 'as', 'much', 'as', 'possible', '.', 'learn', 'more', '.', '.', '.']\n",
            "pred:  0 label:  0 ['images', 'shows', 'how', 'indian', 'police', 'have', 'beaten', 'up', 'migrant', 'workers', 'amid', 'lockdown', 'for', 'commuting']\n",
            "pred:  1 label:  1 ['sage', 'withdraws', 'lockdown', 'claim', '-', 'as', 'uk', 'records', 'highest', 'daily', 'cases', 'since', 'may', '.', 'follow', 'events', 'live', '.']\n",
            "pred:  1 label:  1 ['recent', 'studies', 'show', 'that', 'a', 'significant', 'portion', 'of', 'individuals', 'with', 'lack', 'symptoms', 'and', 'that', 'even', 'those', 'who', 'eventually', 'develop', 'symptoms', 'can', 'transmit', 'virus', 'to', 'others', 'before', 'showing', 'symptoms', '.']\n",
            "pred:  0 label:  0 ['a', 'facebook', 'claims', 'is', 'man', 'made', 'in', 'a', 'lab', '.']\n",
            "pred:  1 label:  1 ['in', 'general', 'however', 'testing', 'situation', 'in', 'united', 'states', 'appears', 'to', 'be', 'deteriorating', 'as', 'infections', 'grow', 'report', 'our', 'colleagues', 'at', '.']\n",
            "pred:  0 label:  0 ['learn', 'from', 'italy', '.', 'no', 'more', 'spaces', 'are', 'available', 'in', 'hospitals', '.', 'please', 'do', 'not', 'step', 'out', 'of', 'your', 'homes', 'for', 'next', '21', 'days', '.']\n",
            "pred:  0 label:  0 ['last', 'year', ',', 'in', 'china', ',', 'a', 'vaccine', 'was', 'compulsory', 'and', 'that', 'vaccine', 'supposedly', 'hosts', 'a', 'digitized', 'virus', ',', 'is', 'one', 'that', 'causes', 'and', 'can', 'be', 'activated', 'through', '5g', 'antennas', '.']\n",
            "pred:  1 label:  1 ['u', '.', 's', '.', 'have', 'enough', 'monkeys', 'for', 'vaccine', 'testing', 'and', 'china', 'has', 'halted', 'exports', '.', '.']\n",
            "pred:  0 label:  0 ['19', 'run', 'away', 'from', 'doctors', 'and', 'police', 'infront', 'of', 'in', 'telangana', '.']\n",
            "pred:  1 label:  1 ['245', 'new', 'cases', 'of', ';', '2802', 'confirmed', 'cases', 'of', 'in', 'nigeria', '417', '93', '.']\n",
            "pred:  1 label:  1 ['testing', 'status', 'stated', 'that', 'samples', 'tested', 'up', 'to', 'september', '05', '2020', 'samples', 'tested', 'on', 'september', '05', '2020', '.']\n",
            "pred:  1 label:  1 ['social', 'distancing', 'works', 'for', 'preventing', 'other', 'infectious', 'diseases', 'too', '.', '.']\n",
            "pred:  1 label:  1 ['we', 'tally', 'and', 'doublecheck', 'state', 'public', 'health', 'numbers', 'with', 'humans', '.', 'we', 'enrich', 'that', 'data', 'with', 'own', 'and', 'other', 'media', '.', 'we', 'do', 'our', 'best', 'to', 'provide', 'standardized', 'numbers', 'but', 'state', 'reporting', 'varies', 'tremendously', 'so', 'we', 'provide', 'maximal', 'transparency', 'into', 'our', 'decisionmaking', '.']\n",
            "pred:  0 label:  0 ['pakistan', 'prime', 'minister', 'imran', 'wife', 'has', 'been', 'tested', 'positive', 'for', 'coronavirus']\n",
            "pred:  0 label:  0 ['a', 'publication', 'that', 'affirms', 'coronavirus', 'was', 'created', 'in', 'united', 'states', 'to', 'kill', 'elderly', 'people', 'and', 'stop', '5g', 'development', '.']\n",
            "pred:  0 label:  0 ['rt', 'had', 'a', 'great', 'time', 'moderating', 'this', 'and', 'misinformation', 'q', 'and', 'a', 'with', 'and']\n",
            "pred:  1 label:  1 ['recovery', 'rate', 'improves', 'to', '77', '.', 'as', 'on', 'september', '05', '2020', 'improvement', 'in', 'recovery', 'rate', 'since', 'initiation', 'on', 'march', '25', '2020', 'via', '.']\n",
            "pred:  0 label:  0 ['online', 'forum', 'website', 'claiming', 'that', '14', 'of', '25', 'confirmed', 'cases', 'of', 'announced', 'on', 'may', '7', 'are', 'oromo', 'political', 'prisoners', '.']\n",
            "pred:  0 label:  0 ['a', 'list', 'of', 'precautions', 'has', 'been', 'issued', 'by', 'indian', 'council', 'of', 'medical', 'research', 'to', 'stay', 'safe', 'from', 'virus', '.']\n",
            "pred:  0 label:  0 ['corona', '?', '?', '?', 'let', 'ppl', 'die', 'i', 'need', 'a', 'photographer', '.', 'well', 'priorities', '.', '.']\n",
            "pred:  1 label:  1 ['pop-up', 'centre', 'in', 'queenstown', 'had', 'over', '1050', 'people', 'showing', 'up', 'on', 'tuesday', 'alone', 'for', 'testing', '.', 'there', 'were', '1076', 'swabs', 'in', 'total', 'taken', 'at', 'pop-up', 'centres', 'all', 'had', 'a', 'negative', 'result', '.']\n",
            "pred:  0 label:  0 ['trump', 'offers', 'to', 'pardon', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['claim', 'saying', 'bill', 'gates', 'has', 'said', 'microchips', 'will', 'be', 'mandatory', 'for', 'travelling', 'after', '.']\n",
            "pred:  0 label:  0 ['says', 'gov', '.', 'tony', 'evers', '?', '?', '?', 'removed', 'american', 'flag', 'from', 'capitol', 'building', '?', '?']\n",
            "pred:  0 label:  0 ['plague', '.']\n",
            "pred:  0 label:  0 ['video', 'shows', 'police', 'forces', 'beating', 'a', 'man', 'on', 'roadside', 'during', 'lockdown', '.']\n",
            "pred:  0 label:  0 ['jimmy', 'kimmel', 'says', 'that', 'broadcasting', 'his', 'show', 'from', 'his', 'home', 'is', 'totally', 'depressing', '.']\n",
            "pred:  0 label:  0 ['black', 'magic', 'at', 'work', 'in', 'bathroom', '.']\n",
            "pred:  1 label:  1 ['rt', 'government', 'on', 'september', '10', '2020', 'released', 'rs', '.', '08', 'crore', 'to', '14', 'states', 'as', 'sixth', 'equated', 'monthly', 'instalment']\n",
            "pred:  0 label:  0 ['during', 'president', 'speech', 'he', 'pointed', 'to', 'a', 'strong', 'economy', 'secure', 'borders', 'and', 'his', 'coronavirus', 'response', 'as', 'reasons', 'why', 'americans', 'should', 're-elect', 'him', 'in', 'november', '.', 'we', '.', '.']\n",
            "pred:  1 label:  1 ['us', 'has', 'now', 'completed', 'tests', 'on', 'over', '1', 'million', 'to', 'be', 'exact', '.', 'a', 'milestone', '.', 'our', 'daily', 'number', 'of', 'completed', 'tests', 'does', 'not', 'appear', 'to', 'be', 'rising', 'anymore', '.', 'for', 'full', 'details', '.', '.']\n",
            "pred:  1 label:  1 ['ncdc', 'has', 'deployed', 'teams', 'to', 'support', 'implementation', 'of', 'response', 'activities', 'funded', 'by', 'project', 'in', 'states', '.', 'teams', 'will', 'provide', 'technical', 'support', 'to', 'individual', 'states', '.', '.']\n",
            "pred:  0 label:  0 ['summer', 'will', 'kill', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['getz', 'pharma', 'imported', '1', ',', '000', 'test', 'kits', '.', 'closed', 'their', 'factory', 'for', '5', 'days', 'and', 'randomly', 'tested', '1', ',', '000', 'out', 'of', 'which', 'were', 'positive', '.', 'these', 'patients', 'were', 'asymptomatic', '.']\n",
            "pred:  0 label:  0 ['delhi', 'public', 'school', 'selling', 'face', 'masks', 'for', 'rs', '.', '400', 'each']\n",
            "pred:  1 label:  1 ['emotional', 'toll', 'of', 'is', 'real', 'especially', 'among', 'some', 'people', 'of', 'racial', 'and', 'ethnic', 'minority', 'groups', 'who', 'have', 'been', 'unequally', 'affected', 'by', 'this', 'pandemic', '.', 'help', 'save', 'a', 'life', '.', '.', '.']\n",
            "pred:  1 label:  1 ['among', 'pregnant', 'women', 'hospitalized', 'for', 'treatment', 'of', '3', 'in', '10', 'required', 'intensive', 'care', 'and', 'one', 'pregnant', 'woman', 'died', 'from', 'a', 'new', 'report', 'of', '105', 'pregnant', 'women', 'in', 'finds', '.', 'find', 'out', '.', '.', '.']\n",
            "pred:  0 label:  0 ['headline', 'claims', 'that', 'macedonian', 'minister', 'of', 'health', 'after', 'a', 'party', 'order', 'of', 'president', 'of', 'sdsm', 'zoran', 'prepared', 'a', 'protocol', 'for', 'election', 'campaign', 'in', 'middle', 'of', 'coronavirus', 'crisis', '.']\n",
            "pred:  0 label:  0 ['muslim', 'people', 'in', 'china', 'seeking', 'refuge', 'in', 'islam', 'by', 'reading', 'quran', 'to', 'save', 'themselves', 'from', '.']\n",
            "pred:  0 label:  0 ['due', 'to', ',', 'would-be', 'robbers', 'have', 'trouble', 'distinguishing', 'themselves', 'from', 'face', 'customers', '.']\n",
            "pred:  0 label:  0 ['there', 'are', 'minors', 'because', 'of', 'in', 'infanta', 'sofía', 'hospital', 'in', 'madrid', 'and', 'three', 'in', '.']\n",
            "pred:  0 label:  0 ['collective', 'on', 'receptors', 'kills', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['washington', 'and', 'new', 'york', 'are', 'driving', 'numbers', '.', 'falling', 'further', 'behind', 'on', 'ca', 'testing', 'numbers', '.', 'now', 'that', 'ma', 'is', 'out', 'in', 'open', 'other', 'big', 'question', 'mark', 'is', 'tx', '.']\n",
            "pred:  1 label:  1 ['there', 'is', 'no', 'one', 'in', 'new', 'zealand', 'receiving', 'care', 'for', '.', 'yesterday', 'our', 'laboratories', 'completed', 'tests', '.', 'that', 'brings', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '.']\n",
            "pred:  0 label:  0 ['data', 'from', '.', 'com', 'shows', 'a', 'massive', 'release', 'of', 'sulfur', 'dioxide', 'gas', 'from', 'outskirts', 'of', 'wuhan', ',', 'commonly', 'associated', 'with', 'burning', 'of', 'organic', 'matters', '.', 'levels', 'are', 'elevated', ',', 'even', 'compared', 'with', 'rest', 'of', 'china']\n",
            "pred:  1 label:  1 ['kia', 'ora', 'vincent', 'to', 'be', 'considered', 'a', 'cluster', 'there', 'must', 'be', 'ten', 'or', 'more', 'cases', 'connected', 'through', 'transmission', 'and', 'who', 'are', 'not', 'all', 'part', 'of', 'same', 'household', '.', 'cluster', 'includes', 'both', 'confirmed', 'and', 'probable', 'cases', '.']\n",
            "pred:  1 label:  1 ['looking', 'to', 'make', 'a', 'splash', 'this', 'summer', '?', 'here', 'are', 'some', 'tips', 'to', 'help', 'prevent', 'spread', 'of', 'for', 'those', 'who', 'operate', 'public', 'pools', 'hot', 'tubs', 'or', 'water', '.', '.']\n",
            "pred:  0 label:  0 ['i', 'have', 'a', 'better', 'give', 'people', 'coronavirus', 'economic', 'relief', 'and', 'hold', 'their', 'hard-earned', 'benefits', 'hostage', '.', '.']\n",
            "pred:  1 label:  1 ['coca', 'call', 'on', 'wednesday', 'june', '3rd', 'at', '2pm', 'et', 'will', 'also', 'be', 'available', 'on', 'facebook', 'profile', 'via', 'facebook', 'live', '.', 'guidance', 'for', 'dental', 'settings', 'during', 'response', '.', 'learn', '.', '.', '.']\n",
            "pred:  0 label:  0 ['in', 'south', 'florida', '?', '?', '?', 'all', 'children', 'and', 'adults', 'must', 'remain', 'in', 'their', 'homes', '?', '?', 'due', 'to', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['a', 'picture', 'with', 'a', 'quote', 'attributed', 'to', 'nobel', 'prize', 'where', 'he', 'says', 'that', 'he', 'was', 'obliged', 'to', 'classify', 'all', 'recent', 'demises', 'as', 'related', 'deaths', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'states', 'reported', 'tests', 'new', 'cases', 'and', '1024', 'deaths', '.', 'number', 'of', 'people', 'hospitalized', 'dropped', 'under', '40000', 'for', 'first', 'time', 'since', 'july', '6', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'deaths', 'per', 'million', 'population', 'in', 'india', 'are', '49', 'as', 'compared', 'to', 'world', 'average', 'of', '111', 'deaths', 'per', 'million', '-', 'secretary']\n",
            "pred:  0 label:  0 ['infamous', 'dubliner', 'janey', 'mac', 'tests', 'positive', 'for', '.']\n",
            "pred:  1 label:  1 ['on', 'total', 'cases', 'in', '5', 'most', 'affected', 'states', 'total', 'number', 'of', 'active', 'cases', 'in', 'country', 'stands', 'at', 'as', 'on', 'date', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'there', 'are', 'simple', 'actions', 'we', 'can', 'all', 'take', 'to', 'prevent', 'spread', 'of', '.', '.']\n",
            "pred:  1 label:  1 ['we', 'continue', 'to', 'progress', 'our', 'contact', 'with', 'people', 'who', 'left', 'managed', 'isolation', 'facilities', 'between', 'june', '9', 'and', 'june', '16', '.', 'we', 'have', 'already', 'followed', 'up', 'with', 'all', '55', 'people', 'who', 'left', 'our', 'facilities', 'under', 'a', 'compassionate', 'exemption', '.']\n",
            "pred:  0 label:  0 ['case', 'count', 'by', 'orlando', 'health', 'in', 'florida', '.', '\"']\n",
            "pred:  0 label:  0 ['a', 'case', 'has', 'been', 'diagnosed', 'in', 'umraniye', 'hospital', 'istanbul', '.']\n",
            "pred:  1 label:  1 ['79', 'new', 'cases', 'of', ';', 'akwa', 'confirmed', 'discharged', '1082', 'deaths', '.']\n",
            "pred:  1 label:  1 ['your', 'look', 'at', 'new', 'cases', 'around', 'nation', '.', 'overall', 'trends', 'changed', 'much', 'this', 'week', '.', '.']\n",
            "pred:  0 label:  0 ['indian', 'army', 'has', 'constructed', 'an', 'advanced', 'hospital', 'with', 'a', 'capacity', 'of', '1000', 'beds', 'in', 'barmer', 'rajasthan', 'to', 'cope', 'with', '.']\n",
            "pred:  1 label:  1 ['california', 'data', 'is', 'all', 'over', 'place', '.', 'they', 'do', 'have', 'a', 'lot', 'of', 'pending', 'tests', 'included', 'in', 'our', 'filtering', 'through', 'to', 'completed', 'tests', 'really', 'slowly', '.']\n",
            "pred:  0 label:  0 ['jeremy', 'corbyn', 'demands', 'government', 'open', 'talks', 'with', 'coronavirus', '.', '.']\n",
            "pred:  0 label:  0 ['man', 'faces', 'impossible', 'choice', 'between', 'protecting', 'himself', 'from', ';', 'looking', 'like', 'hardy', 'bastard', '.']\n",
            "pred:  0 label:  0 ['donating', 'blood', 'can', 'result', 'in', 'testing']\n",
            "pred:  0 label:  0 ['social', 'media', 'posts', 'on', 'are', 'punishable', 'offence', 'now', '.']\n",
            "pred:  1 label:  1 ['new', 'report', 'shows', 'how', 'face', 'coverings', 'can', 'help', 'prevent', '.', '2', 'hair', 'stylists', 'with', 'exposed', 'over', '100', 'clients', '.', 'all', 'wore', 'face', 'coverings', 'and', 'no', 'clients', 'were', 'known', 'to', 'be', 'infected', 'with', '.', 'learn', '.', '.', '.']\n",
            "pred:  1 label:  1 ['number', 'of', 'people', 'in', 'uk', 'who', 'have', 'tested', 'positive', 'for', 'has', 'increased', 'by', '3330', 'in', '24', 'hours', 'official', 'figures', 'show', '.', 'more', 'on', 'this', 'breaking', 'story', '.', '.']\n",
            "pred:  1 label:  1 ['10/10', 'in', 'london', 'is', 'still', 'a', 'mess', '.', '”', 'mayor', 'of', 'london', 'sadiq', 'khan', 'explains', 'what', 'new', 'measures', 'he', 'would', 'like', 'to', 'see', 'in', 'capital', 'and', 'calls', 'to', 'sit', 'in', 'on', 'of', 'cobra', 'meeting', '.', 'get', 'more', 'on', 'plan', 'here', '.', '.']\n",
            "pred:  1 label:  1 ['in', 'new', 'completed', 'tests', 'has', 'leveled', 'off', 'a', 'bit', 'after', 'big', 'mid-week', 'jumps', '.', '.']\n",
            "pred:  1 label:  1 ['recoveries', 'have', 'been', 'registered', 'in', 'last', '24', 'hours', '.', 'centre', 'has', 'been', 'regularly', 'interacting', 'with', 'states', 'that', 'are', 'showing', 'higher', 'fatality', 'rates', '.', 'details', ':', '.']\n",
            "pred:  1 label:  1 ['our', 'daily', 'update', 'is', 'published', '.', 'now', 'tracked', '6', '.', '2', 'million', 'tests', 'up', 'more', 'than', '200k', 'from', 'yesterday', 'in', 'line', 'with', 'last', 'week', '.', 'note', 'that', 'we', 'can', 'only', 'track', 'tests', 'that', 'a', 'state', 'reports', '.', 'for', 'details', '.', '.']\n",
            "pred:  0 label:  0 ['a', 'photo', 'has', 'been', 'shared', 'hundreds', 'of', 'times', 'on', 'facebook', 'alongside', 'a', 'claim', 'it', 'shows', 'coffins', 'of', 'coronavirus', 'victims', 'in', 'italy', '.']\n",
            "pred:  0 label:  0 ['chinese', 'spies', 'stole', 'coronavirus', 'from', 'a', 'canadian', 'lab', '.']\n",
            "pred:  1 label:  1 ['recovery', 'rate', 'improves', 'to', '76', '.', 'as', 'on', 'august', '30', '2020', 'improvement', 'in', 'recovery', 'rate', 'since', 'initiation', 'on', 'march', '25', '2020', 'via', '.']\n",
            "pred:  1 label:  1 ['eight', 'weeks', 'into', 'outbreak', 'and', 'virus', 'is', 'identified', 'sequenced', 'and', 'we', 'have', 'pcr', 'and', 'serological', 'assays', 'are', 'in', 'unprecedented', 'wealth', 'of', 'knowledge', 'for', 'a', 'new', 'disease', '.', 'countries', 'are', 'encouraged', 'to', '.']\n",
            "pred:  1 label:  1 ['people', 'in', 'scotland', 'have', 'been', 'banned', 'from', 'visiting', 'other', 'households', 'indoors', 'in', 'tough', 'new', 'restrictions', 'that', 'go', 'further', 'than', 'those', 'announced', 'for', 'england', '.', 'get', 'more', 'on', 'new', 'measures', 'in', 'scotland', '.', '.']\n",
            "pred:  0 label:  0 ['if', 'you', 'can', 'hold', 'your', 'breath', 'for', '10', 'seconds', 'without', 'coughing', 'or', 'chest', 'pain', 'you', 'are', 'good', 'you', 'have', 'and', 'if', 'your', 'nose', 'is', 'clogged', 'or', 'runny', 'just', 'a', 'cold', '.']\n",
            "pred:  0 label:  0 ['chinese', 'laboratories', 'identified', 'a', 'mystery', 'virus', 'as', 'a', 'highly', 'infectious', 'new', 'pathogen', 'by', 'late', 'december', ',', 'but', 'they', 'were', 'ordered', 'to', 'stop', 'tests', ',', 'destroy', 'samples', 'and', 'suppress', 'news', ',', 'a', 'chinese', 'media', 'has', 'revealed', '.', '|', '.']\n",
            "pred:  0 label:  0 ['a', 'video', 'where', 'bill', 'gates', 'is', 'stating', 'that', 'vaccines', 'are', 'useful', 'to', 'sterilize', 'and', 'reduce', 'world', 'population', '.']\n",
            "pred:  1 label:  1 ['with', 'new', 'case', 'and', 'four', 'additional', 'recovered', 'cases', 'our', 'total', 'number', 'of', 'active', 'cases', 'is', '79', '.', 'of', 'those', '27', 'are', 'imported', 'cases', 'in', 'facilities', 'and', '52', 'are', 'community', 'cases', '.']\n",
            "pred:  1 label:  1 ['5', 'states', 'contribute', 'of', 'total', 'cases', 'of', 'cases', 'and', 'of', 'total', 'reported', 'in', 'india', '.', '.', '.']\n",
            "pred:  1 label:  1 ['be', 'ready', 'for', 'this', 'wave', 'because', 'it', 'will', 'arrive', '.', 'a', 'doctor', 'from', 'spain', 'has', 'warned', 'uk', 'public', 'to', 'wear', 'masks', 'as', 'he', 'says', 'a', 'second', 'wave', 'of', 'is', 'on', 'its', 'way', '.', 'to', 'read', 'more', 'click', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'update', 'on', 'total', 'of', 'samples', 'have', 'been', 'tested', 'till', 'now', '.', 'as', 'reported', 'earlier', 'only', '3', 'were', 'found', 'positive', 'in']\n",
            "pred:  1 label:  1 ['rt', 'are', 'you', 'concerned', 'about', 'possible', 'symptoms', 'and/or', 'exposure', '?', 'check', 'out', 'screening', 'tool', 'developed']\n",
            "pred:  1 label:  1 ['rt', 'going', 'out', '?', 'when', 'you', 'hang', 'out', 'with', 'friends', 'to', 'help', 'protect', 'them', 'from', ';', 'ask', 'them', 'to', 'wear', 'a', 'mask', 'to']\n",
            "pred:  0 label:  0 ['saline', 'water', 'kills', 'new', 'coronavirus', '.']\n",
            "pred:  0 label:  0 ['japanese', 'nobel', 'laureate', 'tasuku', 'honjo', 'announced', 'that', 'coronavirus', 'is', 'man-made', '.']\n",
            "pred:  0 label:  0 ['spanish', 'hospitals', 'are', 'injecting', 'mms', 'to', 'patients', 'to', 'cure', 'them', '.']\n",
            "pred:  0 label:  0 ['and', 'virus', 'aur', 'beech', 'me']\n",
            "pred:  1 label:  1 ['rt', 'gap', 'between', 'cases', 'and', 'cases', 'progressively', 'growing', 'wide', '.', 'more', 'than', '3/4', 'of']\n",
            "pred:  0 label:  0 ['viral', 'message', 'shares', 'indian', 'council', 'of', 'medical', 'research', 'latest', 'guidelines', 'on', '.']\n",
            "pred:  0 label:  0 ['supreme', 'court', 'approves', 'assessment', 'scheme', 'for', 'cancelled', 'class', '10', 'and', '12', 'exams', '.', 'so', 'far', 'no', 'related', 'death', 'has', 'been', 'reported', 'from', 'four', 'northeastern', 'sikkim', 'mizoram', 'manipur', 'and', 'nagaland', '.', '.']\n",
            "pred:  0 label:  0 ['says', 'bill', \"o'reilly\", 'wrote', 'a', 'post', 'claiming', 'that', 'coronavirus', 'was', 'created', 'as', 'a', 'bioweapon', 'by', 'chinese', 'government', '.']\n",
            "pred:  1 label:  1 ['goal', 'of', 'is', 'by', 'end', 'of', '2021', 'to', 'deliver', '2', 'billion', 'doses', 'of', 'safe', 'effective', 'vaccines', 'that', 'will', 'be', 'delivered', 'equally', 'to', 'all', 'participating', 'countries', 'proportional', 'to', 'their', 'populations', '.', '.']\n",
            "pred:  1 label:  1 ['how', 'a', 'healthy', '16-year-old', 'boy', 'lost', 'his', 'life', 'to', 'in', 'just', '12', 'days', '.', '.']\n",
            "pred:  0 label:  0 ['president', 'uhuru', 'kenyatta', 'of', 'kenya', 'ordered', 'credit', 'reference', 'bureaus', 'to', 'delist', 'kenyans', 'who', 'had', 'defaulted', 'on', 'loans', 'to', 'protect', 'kenyans', 'from', 'economic', 'effects', 'of', '.']\n",
            "pred:  0 label:  0 ['president', 'barack', 'obama', 'medical', 'appliance', 'tax', 'bill', 'that', 'forced', 'companies', 'to', 'outsource', 'manufacturing', 'of', 'masks', 'gowns', 'gloves', 'and', 'to', 'china', 'europe', 'and', 'russia', 'to', 'avoid', 'tax', '.', '”']\n",
            "pred:  1 label:  1 ['death', 'toll', 'continues', 'to', 'be', 'concentrated', 'in', 'new', 'york', 'and', 'new', 'jersey', 'although', 'michigan', 'illinois', 'and', 'pennsylvania', 'continue', 'to', 'see', 'substantial', 'numbers', 'of', 'deaths', 'each', 'day', '.', '.']\n",
            "pred:  1 label:  1 ['healthcare', 'workers', 'of', 'color', '2x', 'as', 'likely', 'as', 'whites', 'to', 'get', '.', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'cases', 'are', 'growing', 'exponentially', 'in', 'san', 'bernardino', 'county', 'california', '.', 'yet', 'some', 'are', 'planning', 'to', 'reopen', 'in', 'june', '.', '…']\n",
            "pred:  0 label:  0 ['says', 'new', 'york', 'times', 'exposed', 'reason', 'behind', 'coronavirus', 'crash', 'market', 'to', 'harm', 're-election', 'chances', '.', '”']\n",
            "pred:  0 label:  0 ['in', 'north', 'carolina', ',', '?', '?', '?', 'you', 'can', 'be', 'charged', 'with', 'a', 'class', 'h', 'felony', 'for', 'wearing', 'a', 'mask', 'and', 'concealed', 'carrying', '.', '?', '?']\n",
            "pred:  0 label:  0 ['finally', 'a', 'indian', 'student', 'from', 'pondicherry', 'university', ',', 'named', 'ramu', 'found', 'a', 'home', 'remedy', 'cure', 'for', 'which', 'is', 'for', 'very', 'first', 'time', 'accepted', 'by', 'who', '.']\n",
            "pred:  1 label:  1 ['since', 'august', '12', 'our', 'contact', 'tracing', 'team', 'has', 'identified', 'close', 'contacts', 'of', 'cases', 'of', 'which', 'have', 'been', 'contacted', 'and', 'are', 'and', 'we', 'are', 'in', 'process', 'of', 'contacting', 'rest', '.']\n",
            "pred:  0 label:  0 ['a', 'bio-engineered', 'pandemic', 'is', 'too', 'useful', 'to', 'waste', '.', '.', '.', 'lockdowns', ',', 'new', 'gulags', '.']\n",
            "pred:  0 label:  0 ['exposed', 'to', 'sun', 'for', 'two', 'kills', '2019', 'coronavirus', '.']\n",
            "pred:  1 label:  1 ['did', 'you', 'know', '?', 'among', 'adults', 'risk', 'of', 'severe', 'illness', 'from', 'increases', 'with', 'age', '.', 'learn', 'steps', 'to', 'reduce', 'your', '.', '.']\n",
            "pred:  0 label:  0 ['imagine', 'how', 'much', 'better', 'off', 'we', 'would', 'be', 'if', 'a', 'had', 'been', 'mantra', 'from', 'day', '1', '.', 'searches', 'to', 'buy', 'chloroquine', 'spiked', 'by', 'following', 'donald', 'trump', 'and', 'elon', 'endorsements', 'of', 'drug', 'for', 'treating']\n",
            "pred:  0 label:  0 ['mosquitoes', 'can', 'transfer', '?', '?', '?', 'from', 'person', 'to', 'person', '?', '?']\n",
            "pred:  1 label:  1 ['return', 'of', 'higher', 'testing', 'numbers', 'has', 'been', 'driven', 'not', 'just', 'by', 'south', 'and', 'west', 'but', 'also', 'northeast', 'which', 'still', 'has', 'a', 'much', 'lower', 'case', 'load', '.', '7-day', 'average', 'high', 'is', 'marked', 'for', 'each', 'region', '.', ')', '.']\n",
            "pred:  1 label:  1 ['when', 'parents', 'and', 'caregivers', 'communicate', 'about', 'with', 'their', 'children', 'calmly', 'and', 'openly', 'they', 'can', 'lessen', 'their', 'fears', 'about', 'disease', '.', 'learn', 'more', 'about', 'talking', 'with', 'children', 'about', '.', '.', '.']\n",
            "pred:  1 label:  1 ['after', 'running', 'near', 'for', 'most', 'of', 'april', 'cumulative', 'positive', 'rate', 'has', 'fallen', 'a', 'couple', 'percent', '.', 'all', 'of', 'these', 'things', 'are', 'good', 'signs', 'even', 'if', 'we', 'did', 'add', 'more', 'than', 'positive', 'cases', 'to', 'rolls', '.']\n",
            "pred:  0 label:  0 ['?', '?', '?', 'michigan', 'governor', 'gretchen', 'whitmer', 'bans', 'buying', 'us', 'flags', 'during', 'lockdown', '.', '?', '?']\n",
            "pred:  0 label:  0 ['people', 'having', 'a', 'beer', 'on', 'beach', 'to', 'blame', 'for', 'mishandling', 'of', 'coronavirus', 'crisis', '.', '.']\n",
            "pred:  0 label:  0 ['putting', 'little', 'bastards', 'up', 'for', 'read', 'some', 'of', 'heartwarming', 'stories', '.']\n",
            "pred:  0 label:  0 ['u', '.', 's', '.', 'coronavirus', 'deaths', 'top', '1100', 'for', 'a', 'third', 'day', 'in', 'a', 'row', '.']\n",
            "pred:  0 label:  0 ['cases', 'are', 'not', 'spiking', 'and', 'they', 'have', 'no', 'face', 'masks', 'and', 'no', 'social', 'distancing', '.', 'and', 'if', 'you', 'get', ',', 'you', 'just', 'have', 'to', 'take', ';', 'zinc', 'and', 'gone', 'in', '6', 'days', '.']\n",
            "pred:  1 label:  1 ['second', 'wave', 'would', 'happen', 'says', 'pm', '.', 'boris', 'johnson', 'has', 'said', 'that', 'he', 'want', 'a', 'second', 'lockdown', 'but', 'will', 'consider', 'if', 'current', 'measures', 'to', 'go', '.', 'read', 'more', '.', '.']\n",
            "pred:  0 label:  0 ['inhaling', 'hot', 'water', 'steam', 'kills', 'corona', 'virus', '-', 'homemade']\n",
            "pred:  1 label:  1 ['total', 'deaths', 'reached', '.', 'states', 'reported', '1904', 'deaths', 'today', 'about', 'same', 'level', 'as', 'past', 'two', 'days', '.', '.']\n",
            "pred:  1 label:  1 ['yesterday', 'our', 'laboratories', 'completed', '800', 'tests', 'bringing', 'total', 'number', 'of', 'tests', 'completed', 'to', 'date', 'to', '-', 'lower', 'testing', 'volumes', 'are', 'regularly', 'observed', 'over', 'weekends', '.']\n",
            "pred:  0 label:  0 ['everyone', 'knows', 'that', 'facing', 'a', 'real', 'crisis', 'from', 'coronavirus', '.', 'but', 'do', 'you', 'know', 'how', 'we', 'got', 'here', 'and', 'what', 'we', 'need', 'to', 'do', 'next', '?', 'ron', 'klain', ',', 'former', 'white', 'house', 'ebola', 'response', 'coordinator', ',', 'breaks', 'it', 'down', 'for', '.']\n",
            "pred:  0 label:  0 ['hospital', 'flooded', 'in', 'bihar']\n",
            "pred:  1 label:  1 ['contact', 'tracing', 'aims', 'to', 'identify', 'and', 'alert', 'people', 'who', 'have', 'come', 'into', 'contact', 'with', 'a', 'person', 'infected', 'with', '.', 'experts', 'recommend', 'tracing', 'contacts', 'of', 'someone', 'who', 'tests', 'positive', 'for', 'within', '48', 'hours', 'to', 'contain', 'potential', 'of', 'transmission', '.']\n",
            "pred:  1 label:  1 ['current', 'evidence', 'shows', 'that', 'about', 'of', 'patients', 'experience', 'mild', 'symptoms', ';', 'recover', 'ncdc', 'recommends', 'treatment', 'of', 'every', 'positive', 'case', 'in', 'an', 'isolation', 'centre', 'to', 'mitigate', 'risk', 'of', 'spread', 'in', 'communities', '.']\n",
            "pred:  0 label:  0 ['numbers', 'are', 'better', 'than', 'almost', 'all', 'countries', '.', '\"']\n",
            "pred:  1 label:  1 ['11', 'new', 'cases', 'of', 'have', 'been', 'reported', 'in', '8', 'in', 'lagos', '2', 'in', 'enugu', ';', '1', 'in', 'edo', 'state', 'as', 'at', '27th', 'march', 'there', 'are', '81', 'confirmed', 'cases', 'of', 'reported', 'in', 'nigeria', '.', '3', 'have', 'been', 'discharged', 'with', '1', 'death', '.', '.']\n",
            "pred:  0 label:  0 ['banana', 'contains', 'a', 'lectin', 'that', 'is', 'a', 'powerful', 'agent', 'blocking', 'virus', 'from', 'entering', 'cell', '.', 'eating', 'a', 'banana', 'a', 'day', 'is', 'recommended', '.']\n",
            "pred:  1 label:  1 ['asymptomatic', 'coronavirus', 'spread', 'is', 'rare', 'says', '.', '.']\n",
            "pred:  1 label:  1 ['auckland', 'is', 'to', 'remain', 'at', 'alert', 'level', '2', 'with', 'restrictions', 'on', 'gatherings', '2', '.', 'until', 'at', 'least', '14', 'september', '.', 'social', 'gatherings', 'in', 'auckland', 'are', 'limited', 'to', 'no', 'more', 'than', '10', 'people', 'with', '50', 'allowed', 'for', 'authorised', 'funerals', 'and', '.', 'rest', 'of', 'country', 'is', 'at', 'level', '2', '.']\n",
            "pred:  1 label:  1 ['lagos', 'state', 'recorded', '25', 'deaths', 'between', 'june', '12th', ';', '15th', ';', 'they', 'were', 'all', 'announced', 'on', 'june', '16th', 'a', 'breakdown', 'of', 'cases', 'by', 'state', 'can', 'be', 'found', 'via', '.', '.']\n",
            "pred:  1 label:  1 ['over', '1500', 'deaths', 'were', 'reported', '.', '7-day', 'average', 'fell', 'under', '1300', 'deaths', 'for', 'first', 'time', 'since', 'april', '7', '.', 'still', 'nation', 'is', 'likely', 'to', 'pass', '100000', 'probable', 'and', 'confirmed', 'deaths', 'within', 'a', 'week', '.', '.']\n",
            "pred:  0 label:  0 ['coronavirus', 'crisis', 'in', 'italy', 'leads', 'donald', 'trump', 'to', 'close', 'all', 'us', 'pizzerias', '.', '.']\n",
            "pred:  1 label:  1 ['new', 'measures', 'have', 'been', 'introduced', 'by', 'spanish', 'government', 'as', 'infection', 'rates', 'exceed', 'one', 'in', '100', 'people', 'in', 'some', 'of', 'areas', 'in', 'city', 'and', 'its', 'outskirts', 'home', 'to', 'around', 'people', '.']\n",
            "pred:  1 label:  1 ['this', 'paradigm', 'shift', 'a', 'how', 'behind', 'inpatient', 'diabetes', 'management', 'has', 'been', '.', '.']\n",
            "pred:  1 label:  1 ['rt', 'virus', 'been', 'found', 'to', 'date', 'in', 'breast', 'benefits', 'of', 'outweigh', 'potential']\n",
            "pred:  0 label:  0 ['alkaline', 'food', 'can', 'cure']\n",
            "pred:  0 label:  0 ['tonight', 'onwards', 'disaster', 'management', 'act', 'has', 'been', 'implemented', 'across', 'country', '.', 'according', 'to', 'this', 'update', ',', 'apart', 'from', 'govt', 'department', 'no', 'other', 'citizen', 'is', 'allowed', 'to', 'post', 'any', 'update', 'or', 'share', 'any', 'forward', 'related', 'to', 'coronavirus', ',', 'it', 'being', 'punishable', 'offence', '.']\n",
            "pred:  1 label:  1 ['296', 'new', 'cases', 'of', ';', 'akwa', 'confirmed', 'discharged', '1011', 'deaths', '.']\n"
          ]
        }
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "CovidTransformer.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.8.5"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}