messy_vs_clean_room.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "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": "3.5.2"
    },
    "colab": {
      "name": "messy-vs-clean-room.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/WinstonHChou/PANO-Project/blob/master/messy_vs_clean_room.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "18pFn7sBWVQL",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "bf603b39-d624-49e2-9dff-8c2cfd7e26f6"
      },
      "source": [
        "import keras\n",
        "keras.__version__"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'2.3.1'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jQWml8rVWWFE",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "28dfc3dd-0e56-40d6-8215-3b18e14de2d1"
      },
      "source": [
        "import tensorflow as tf\n",
        "print(\"GPU Available: \", tf.config.list_physical_devices('GPU'))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "GPU Available:  [PhysicalDevice(name='/physical_device:GPU:0', device_type='GPU')]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QL_RnTvoWVQR",
        "colab_type": "text"
      },
      "source": [
        "We download the data"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RAt38H4Fmghl",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 445
        },
        "outputId": "14cab120-c804-476c-a533-77700d6df608"
      },
      "source": [
        "FILEID='17BB2Ufj-9rTnT9cwZR8fu_sKKGdCXLxM'\n",
        "FILENAME='train.zip'\n",
        "!wget --load-cookies /tmp/cookies.txt \"https://docs.google.com/uc?export=download&confirm=$(wget --quiet --save-cookies /tmp/cookies.txt --keep-session-cookies --no-check-certificate 'https://docs.google.com/uc?export=download&id=$FILEID' -O- | sed -rn 's/.*confirm=([0-9A-Za-z_]+).*/\\1\\n/p')&id=$FILEID\" -O $FILENAME && rm -rf /tmp/cookies.txt"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2020-07-24 00:24:30--  https://docs.google.com/uc?export=download&confirm=TxK9&id=17BB2Ufj-9rTnT9cwZR8fu_sKKGdCXLxM\n",
            "Resolving docs.google.com (docs.google.com)... 173.194.79.102, 173.194.79.138, 173.194.79.101, ...\n",
            "Connecting to docs.google.com (docs.google.com)|173.194.79.102|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Moved Temporarily\n",
            "Location: https://doc-0g-8g-docs.googleusercontent.com/docs/securesc/thv7qi5llc8bfc68lu5rgoko02l6saj3/4r75ej1tntimcureo0jcv3pfl0i100gs/1595550225000/12308918870841825745/05911354716773686956Z/17BB2Ufj-9rTnT9cwZR8fu_sKKGdCXLxM?e=download [following]\n",
            "--2020-07-24 00:24:31--  https://doc-0g-8g-docs.googleusercontent.com/docs/securesc/thv7qi5llc8bfc68lu5rgoko02l6saj3/4r75ej1tntimcureo0jcv3pfl0i100gs/1595550225000/12308918870841825745/05911354716773686956Z/17BB2Ufj-9rTnT9cwZR8fu_sKKGdCXLxM?e=download\n",
            "Resolving doc-0g-8g-docs.googleusercontent.com (doc-0g-8g-docs.googleusercontent.com)... 74.125.128.132, 2a00:1450:4013:c02::84\n",
            "Connecting to doc-0g-8g-docs.googleusercontent.com (doc-0g-8g-docs.googleusercontent.com)|74.125.128.132|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Found\n",
            "Location: https://docs.google.com/nonceSigner?nonce=i7t5fm1563gis&continue=https://doc-0g-8g-docs.googleusercontent.com/docs/securesc/thv7qi5llc8bfc68lu5rgoko02l6saj3/4r75ej1tntimcureo0jcv3pfl0i100gs/1595550225000/12308918870841825745/05911354716773686956Z/17BB2Ufj-9rTnT9cwZR8fu_sKKGdCXLxM?e%3Ddownload&hash=geedicd04rofn2cl53ri3qdgahh8jm5s [following]\n",
            "--2020-07-24 00:24:31--  https://docs.google.com/nonceSigner?nonce=i7t5fm1563gis&continue=https://doc-0g-8g-docs.googleusercontent.com/docs/securesc/thv7qi5llc8bfc68lu5rgoko02l6saj3/4r75ej1tntimcureo0jcv3pfl0i100gs/1595550225000/12308918870841825745/05911354716773686956Z/17BB2Ufj-9rTnT9cwZR8fu_sKKGdCXLxM?e%3Ddownload&hash=geedicd04rofn2cl53ri3qdgahh8jm5s\n",
            "Connecting to docs.google.com (docs.google.com)|173.194.79.102|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Found\n",
            "Location: https://doc-0g-8g-docs.googleusercontent.com/docs/securesc/thv7qi5llc8bfc68lu5rgoko02l6saj3/4r75ej1tntimcureo0jcv3pfl0i100gs/1595550225000/12308918870841825745/05911354716773686956Z/17BB2Ufj-9rTnT9cwZR8fu_sKKGdCXLxM?e=download&nonce=i7t5fm1563gis&user=05911354716773686956Z&hash=4tip2ouma948rrtdd8eugk2h4ka33dj7 [following]\n",
            "--2020-07-24 00:24:32--  https://doc-0g-8g-docs.googleusercontent.com/docs/securesc/thv7qi5llc8bfc68lu5rgoko02l6saj3/4r75ej1tntimcureo0jcv3pfl0i100gs/1595550225000/12308918870841825745/05911354716773686956Z/17BB2Ufj-9rTnT9cwZR8fu_sKKGdCXLxM?e=download&nonce=i7t5fm1563gis&user=05911354716773686956Z&hash=4tip2ouma948rrtdd8eugk2h4ka33dj7\n",
            "Connecting to doc-0g-8g-docs.googleusercontent.com (doc-0g-8g-docs.googleusercontent.com)|74.125.128.132|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: unspecified [application/zip]\n",
            "Saving to: ‘train.zip’\n",
            "\n",
            "train.zip               [  <=>               ]  32.12M  64.6MB/s    in 0.5s    \n",
            "\n",
            "2020-07-24 00:24:33 (64.6 MB/s) - ‘train.zip’ saved [33681249]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6uyKGqmjZR3G",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        },
        "outputId": "a71cef6d-a90b-40fb-e3fb-a09a54364f55"
      },
      "source": [
        "ls -lh"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "total 34M\n",
            "-rw-r--r-- 1 root root 1.1M Jul 24 00:18 Messy-rooms-make-nice-memories.jpg\n",
            "drwxr-xr-x 1 root root 4.0K Jul 10 16:29 \u001b[0m\u001b[01;34msample_data\u001b[0m/\n",
            "-rw-r--r-- 1 root root  33M Jul 24 00:24 train.zip\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3_95pNVUZa7x",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!unzip -q train.zip -d kaggle_original_data"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E-vcDoYMeftP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "!rm -r kaggle_original_data/__MACOSX/"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "u16kusuQeAQ0",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "20c2e6a6-941f-4a5c-80fc-f00e2e0a0de4"
      },
      "source": [
        "!ls -l kaggle_original_data | head"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "total 8\n",
            "drwxr-xr-x 2 root root 4096 Jul 22 20:17 clean\n",
            "drwxr-xr-x 2 root root 4096 Jul 22 20:17 messy\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WciW0r6LKf1Z",
        "colab_type": "text"
      },
      "source": [
        "We now are going to sort the images by separating them into different folders."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7h2qRQnQWVQR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from random import shuffle\n",
        "import os, shutil\n",
        "\n",
        "# list all labels\n",
        "label_dirs = ['clean', 'messy']\n",
        "\n",
        "# The path to the directory where the original\n",
        "# dataset was uncompressed\n",
        "original_dataset_dir = '/content/kaggle_original_data'\n",
        "\n",
        "# The directory where we will\n",
        "# store our smaller dataset\n",
        "base_dir = '/content/messy_vs_clean_room'\n",
        "os.mkdir(base_dir)\n",
        "\n",
        "# Directories for our training,\n",
        "# validation and test splits\n",
        "train_dir = os.path.join(base_dir, 'train')\n",
        "os.mkdir(train_dir)\n",
        "validation_dir = os.path.join(base_dir, 'validation')\n",
        "os.mkdir(validation_dir)\n",
        "test_dir = os.path.join(base_dir, 'test')\n",
        "os.mkdir(test_dir)\n",
        "\n",
        "# Directory with our training/validation/test label pictures\n",
        "for target_dir in [train_dir, validation_dir, test_dir]:\n",
        "    for label in label_dirs:\n",
        "        dir = os.path.join(target_dir, label)\n",
        "        os.mkdir(dir)\n",
        "\n",
        "# Copy 70% of each label to train, 15% to valid, and 15% to test directories\n",
        "for label in label_dirs:\n",
        "    fnames = os.listdir(os.path.join(original_dataset_dir, label))\n",
        "    shuffle(fnames)  # shuffling the list\n",
        "    n_img_start_valid = int(len(fnames)*0.7)\n",
        "    n_img_start_test  = int(len(fnames)*0.85)\n",
        "    for fname in fnames[:n_img_start_valid]:  # train\n",
        "        src = os.path.join(original_dataset_dir, label, fname)\n",
        "        dst = os.path.join(train_dir, label, fname)\n",
        "        shutil.copyfile(src, dst)\n",
        "    for fname in fnames[n_img_start_valid:n_img_start_test]:  # valid\n",
        "        src = os.path.join(original_dataset_dir, label, fname)\n",
        "        dst = os.path.join(validation_dir, label, fname)\n",
        "        shutil.copyfile(src, dst)\n",
        "    for fname in fnames[n_img_start_test:]:  # test\n",
        "        src = os.path.join(original_dataset_dir, label, fname)\n",
        "        dst = os.path.join(test_dir, label, fname)\n",
        "        shutil.copyfile(src, dst)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lz_EVN0jWVQW",
        "colab_type": "text"
      },
      "source": [
        "As a sanity check, let's count how many pictures we have in each training split (train / validation / test):"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M5S62youWVQX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "cd94ed88-58da-45d2-868e-c2d58e97cfca"
      },
      "source": [
        "total_train_imgs = 0\n",
        "total_valid_imgs = 0\n",
        "for label in label_dirs:\n",
        "    print('total images for label', label, 'in training:', len(os.listdir(os.path.join(train_dir, label))),\n",
        "                                           'in valid:', len(os.listdir(os.path.join(validation_dir, label))),\n",
        "                                            'in test:', len(os.listdir(os.path.join(test_dir, label))))\n",
        "    total_train_imgs += len(os.listdir(os.path.join(train_dir, label)))\n",
        "    total_valid_imgs += len(os.listdir(os.path.join(validation_dir, label)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "total images for label clean in training: 74 in valid: 16 in test: 16\n",
            "total images for label messy in training: 74 in valid: 16 in test: 16\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MKRmuMiuN7PK",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "b141fccb-5f02-4b4d-efa3-1577179c8be8"
      },
      "source": [
        "print('Total number of training images:', total_train_imgs)\n",
        "print('Total number of validation images:', total_valid_imgs)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Total number of training images: 148\n",
            "Total number of validation images: 32\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nILJEVrsWVQj",
        "colab_type": "text"
      },
      "source": [
        "## Building our network\n",
        "\n",
        "We've already built a small convnet for MNIST in the previous example, so you should be familiar with them. We will reuse the same \n",
        "general structure: our convnet will be a stack of alternated `Conv2D` (with `relu` activation) and `MaxPooling2D` layers.\n",
        "\n",
        "However, since we are dealing with bigger images and a more complex problem, we will make our network accordingly larger: it will have one \n",
        "more `Conv2D` + `MaxPooling2D` stage. This serves both to augment the capacity of the network, and to further reduce the size of the \n",
        "feature maps, so that they aren't overly large when we reach the `Flatten` layer. Here, since we start from inputs of size 150x150 (a \n",
        "somewhat arbitrary choice), we end up with feature maps of size 7x7 right before the `Flatten` layer.\n",
        "\n",
        "Note that the depth of the feature maps is progressively increasing in the network (from 32 to 128), while the size of the feature maps is \n",
        "decreasing (from 148x148 to 7x7). This is a pattern that you will see in almost all convnets.\n",
        "\n",
        "Since we are attacking a binary classification problem, we are ending the network with a single unit (a `Dense` layer of size 1) and a \n",
        "`sigmoid` activation. This unit will encode the probability that the network is looking at one class or the other."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2ZZEoAowOBMi",
        "colab_type": "text"
      },
      "source": [
        "Original Model (4 Conv2D + MaxPooling2D layers)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RWWyLW_GDVKo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras import layers\n",
        "from keras import models\n",
        "\n",
        "model = models.Sequential()\n",
        "model.add(layers.Conv2D(32, (3, 3), activation='relu',\n",
        "                        input_shape=(150, 150, 3)))\n",
        "model.add(layers.MaxPooling2D((2, 2)))\n",
        "model.add(layers.Conv2D(64, (3, 3), activation='relu'))\n",
        "model.add(layers.MaxPooling2D((2, 2)))\n",
        "model.add(layers.Conv2D(128, (3, 3), activation='relu'))\n",
        "model.add(layers.MaxPooling2D((2, 2)))\n",
        "model.add(layers.Conv2D(128, (3, 3), activation='relu'))\n",
        "model.add(layers.MaxPooling2D((2, 2)))\n",
        "model.add(layers.Flatten())\n",
        "model.add(layers.Dense(512, activation='relu'))\n",
        "model.add(layers.Dense(1, activation='sigmoid'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sQUnFlv9EOMM",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rRp6ROwoERx1",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 232
        },
        "outputId": "a38bd568-488e-4978-9b82-0ee66ec915ed"
      },
      "source": [
        "from keras import optimizers\n",
        "\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.RMSprop(lr=1e-4),\n",
        "              metrics=['acc'])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-44-4abcf84ae7e0>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0mkeras\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0moptimizers\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m model.compile(loss='binary_crossentropy',\n\u001b[0m\u001b[1;32m      4\u001b[0m               \u001b[0moptimizer\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0moptimizers\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mRMSprop\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlr\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m1e-4\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m               metrics=['acc'])\n",
            "\u001b[0;31mNameError\u001b[0m: name 'model' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ILkVXqSNEVrD",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "3cf307cd-fb00-4b04-d135-dbcb82681853"
      },
      "source": [
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "\n",
        "# All images will be rescaled by 1./255\n",
        "train_datagen = ImageDataGenerator(rescale=1./255)\n",
        "test_datagen = ImageDataGenerator(rescale=1./255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "        # This is the target directory\n",
        "        train_dir,\n",
        "        # All images will be resized to 150x150\n",
        "        target_size=(150, 150),\n",
        "        batch_size=20,\n",
        "        # Since we use binary_crossentropy loss, we need binary labels\n",
        "        class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "        validation_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=20,\n",
        "        class_mode='binary')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 148 images belonging to 2 classes.\n",
            "Found 32 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kEAuJfIlEfk6",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "d724085b-1968-444a-a021-dcfdbf4b8864"
      },
      "source": [
        "for data_batch, labels_batch in train_generator:\n",
        "    print('data batch shape:', data_batch.shape)\n",
        "    print('labels batch shape:', labels_batch.shape)\n",
        "    break"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "data batch shape: (20, 150, 150, 3)\n",
            "labels batch shape: (20,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d1Io5Tz0EpBg",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "7862aa07-4761-4d0f-cf3a-fb0b4c87fc51"
      },
      "source": [
        "history = model.fit_generator(\n",
        "      train_generator,\n",
        "      steps_per_epoch=100,\n",
        "      epochs=30,\n",
        "      validation_data=validation_generator,\n",
        "      validation_steps=50)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/30\n",
            "100/100 [==============================] - 10s 101ms/step - loss: 0.5283 - acc: 0.7274 - val_loss: 0.4134 - val_acc: 0.7500\n",
            "Epoch 2/30\n",
            "100/100 [==============================] - 11s 107ms/step - loss: 0.3404 - acc: 0.8639 - val_loss: 0.8140 - val_acc: 0.7188\n",
            "Epoch 3/30\n",
            "100/100 [==============================] - 11s 110ms/step - loss: 0.2346 - acc: 0.9143 - val_loss: 0.8827 - val_acc: 0.7500\n",
            "Epoch 4/30\n",
            "100/100 [==============================] - 11s 107ms/step - loss: 0.1304 - acc: 0.9596 - val_loss: 0.8549 - val_acc: 0.7812\n",
            "Epoch 5/30\n",
            "100/100 [==============================] - 11s 105ms/step - loss: 0.0670 - acc: 0.9811 - val_loss: 1.0388 - val_acc: 0.7188\n",
            "Epoch 6/30\n",
            "100/100 [==============================] - 10s 103ms/step - loss: 0.0471 - acc: 0.9870 - val_loss: 0.7218 - val_acc: 0.6875\n",
            "Epoch 7/30\n",
            "100/100 [==============================] - 11s 108ms/step - loss: 0.0490 - acc: 0.9881 - val_loss: 1.4416 - val_acc: 0.8125\n",
            "Epoch 8/30\n",
            "100/100 [==============================] - 11s 114ms/step - loss: 0.0096 - acc: 0.9989 - val_loss: 1.9165 - val_acc: 0.6875\n",
            "Epoch 9/30\n",
            "100/100 [==============================] - 10s 105ms/step - loss: 0.0306 - acc: 0.9881 - val_loss: 3.0693 - val_acc: 0.6562\n",
            "Epoch 10/30\n",
            "100/100 [==============================] - 11s 109ms/step - loss: 0.0155 - acc: 0.9962 - val_loss: 2.5212 - val_acc: 0.6875\n",
            "Epoch 11/30\n",
            "100/100 [==============================] - 10s 99ms/step - loss: 0.0215 - acc: 0.9957 - val_loss: 0.3182 - val_acc: 0.6562\n",
            "Epoch 12/30\n",
            "100/100 [==============================] - 10s 104ms/step - loss: 0.0212 - acc: 0.9962 - val_loss: 0.6931 - val_acc: 0.6562\n",
            "Epoch 13/30\n",
            "100/100 [==============================] - 10s 96ms/step - loss: 1.8876e-04 - acc: 1.0000 - val_loss: 4.5183 - val_acc: 0.5625\n",
            "Epoch 14/30\n",
            "100/100 [==============================] - 10s 100ms/step - loss: 0.0263 - acc: 0.9962 - val_loss: 3.4565 - val_acc: 0.6562\n",
            "Epoch 15/30\n",
            "100/100 [==============================] - 10s 102ms/step - loss: 0.0307 - acc: 0.9951 - val_loss: 4.5342 - val_acc: 0.6562\n",
            "Epoch 16/30\n",
            "100/100 [==============================] - 10s 101ms/step - loss: 0.0197 - acc: 0.9930 - val_loss: 1.5415 - val_acc: 0.7812\n",
            "Epoch 17/30\n",
            "100/100 [==============================] - 9s 94ms/step - loss: 4.6627e-05 - acc: 1.0000 - val_loss: 4.2565 - val_acc: 0.7188\n",
            "Epoch 18/30\n",
            "100/100 [==============================] - 10s 101ms/step - loss: 0.0097 - acc: 0.9957 - val_loss: 4.6585 - val_acc: 0.7812\n",
            "Epoch 19/30\n",
            "100/100 [==============================] - 10s 101ms/step - loss: 0.0206 - acc: 0.9941 - val_loss: 4.4193 - val_acc: 0.6875\n",
            "Epoch 20/30\n",
            "100/100 [==============================] - 10s 96ms/step - loss: 9.9015e-06 - acc: 1.0000 - val_loss: 3.3249 - val_acc: 0.6875\n",
            "Epoch 21/30\n",
            "100/100 [==============================] - 9s 95ms/step - loss: 0.0238 - acc: 0.9946 - val_loss: 7.8197 - val_acc: 0.6875\n",
            "Epoch 22/30\n",
            "100/100 [==============================] - 10s 104ms/step - loss: 7.4188e-05 - acc: 1.0000 - val_loss: 2.9433 - val_acc: 0.6875\n",
            "Epoch 23/30\n",
            "100/100 [==============================] - 9s 95ms/step - loss: 0.0333 - acc: 0.9962 - val_loss: 3.2964 - val_acc: 0.6875\n",
            "Epoch 24/30\n",
            "100/100 [==============================] - 10s 97ms/step - loss: 3.7408e-05 - acc: 1.0000 - val_loss: 0.5912 - val_acc: 0.6562\n",
            "Epoch 25/30\n",
            "100/100 [==============================] - 10s 99ms/step - loss: 8.7447e-07 - acc: 1.0000 - val_loss: 6.0348 - val_acc: 0.6562\n",
            "Epoch 26/30\n",
            "100/100 [==============================] - 10s 104ms/step - loss: 0.0306 - acc: 0.9952 - val_loss: 7.7512 - val_acc: 0.6562\n",
            "Epoch 27/30\n",
            "100/100 [==============================] - 10s 95ms/step - loss: 1.7021e-06 - acc: 1.0000 - val_loss: 1.9042 - val_acc: 0.6562\n",
            "Epoch 28/30\n",
            "100/100 [==============================] - 10s 98ms/step - loss: 0.0064 - acc: 0.9973 - val_loss: 3.2348 - val_acc: 0.7500\n",
            "Epoch 29/30\n",
            "100/100 [==============================] - 9s 93ms/step - loss: 4.5082e-07 - acc: 1.0000 - val_loss: 0.9439 - val_acc: 0.7500\n",
            "Epoch 30/30\n",
            "100/100 [==============================] - 10s 100ms/step - loss: 0.0079 - acc: 0.9973 - val_loss: 2.8037 - val_acc: 0.7188\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AmvUCu48qQQl",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.save('clean_and_messy_1.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hbYe4ULuFz_G",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "outputId": "87a9ac05-b562-489e-c557-ca5374ca937e"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "acc = history.history['acc']\n",
        "val_acc = history.history['val_acc']\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "epochs = range(len(acc))\n",
        "\n",
        "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.legend()\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "97SRjMnTdIBa",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 238
        },
        "outputId": "f4360a7f-5320-4515-e8ae-0cd4660fecad"
      },
      "source": [
        "test_generator = test_datagen.flow_from_directory(\n",
        "        test_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=20,\n",
        "        class_mode='binary')\n",
        "\n",
        "test_loss, test_acc = model.evaluate_generator(test_generator, steps=50)\n",
        "print('test acc:', test_acc)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 32 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-22-19f8443b6c42>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m      5\u001b[0m         class_mode='binary')\n\u001b[1;32m      6\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 7\u001b[0;31m \u001b[0mtest_loss\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtest_acc\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mevaluate_generator\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mtest_generator\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0msteps\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m50\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      8\u001b[0m \u001b[0mprint\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m'test acc:'\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mtest_acc\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'model' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vR4sZ4hwZwYq",
        "colab_type": "text"
      },
      "source": [
        "---"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UEAiUmRtKvcV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "datagen = ImageDataGenerator(\n",
        "      rotation_range=40,\n",
        "      width_shift_range=0.2,\n",
        "      height_shift_range=0.2,\n",
        "      shear_range=0.2,\n",
        "      zoom_range=0.2,\n",
        "      horizontal_flip=True,\n",
        "      fill_mode='nearest')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "td329wQGLFKY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model = models.Sequential()\n",
        "model.add(layers.Conv2D(32, (3, 3), activation='relu',\n",
        "                        input_shape=(150, 150, 3)))\n",
        "model.add(layers.MaxPooling2D((2, 2)))\n",
        "model.add(layers.Conv2D(64, (3, 3), activation='relu'))\n",
        "model.add(layers.MaxPooling2D((2, 2)))\n",
        "model.add(layers.Conv2D(128, (3, 3), activation='relu'))\n",
        "model.add(layers.MaxPooling2D((2, 2)))\n",
        "model.add(layers.Conv2D(128, (3, 3), activation='relu'))\n",
        "model.add(layers.MaxPooling2D((2, 2)))\n",
        "model.add(layers.Flatten())\n",
        "model.add(layers.Dropout(0.5))\n",
        "model.add(layers.Dense(512, activation='relu'))\n",
        "model.add(layers.Dense(1, activation='sigmoid'))\n",
        "\n",
        "model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.RMSprop(lr=2e-5),\n",
        "              metrics=['acc'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AByAllPfTRHK",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 580
        },
        "outputId": "d9615300-6cd2-4d26-f360-6c629941c769"
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_4\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_9 (Conv2D)            (None, 148, 148, 32)      896       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_9 (MaxPooling2 (None, 74, 74, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_10 (Conv2D)           (None, 72, 72, 64)        18496     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_10 (MaxPooling (None, 36, 36, 64)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_11 (Conv2D)           (None, 34, 34, 128)       73856     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_11 (MaxPooling (None, 17, 17, 128)       0         \n",
            "_________________________________________________________________\n",
            "conv2d_12 (Conv2D)           (None, 15, 15, 128)       147584    \n",
            "_________________________________________________________________\n",
            "max_pooling2d_12 (MaxPooling (None, 7, 7, 128)         0         \n",
            "_________________________________________________________________\n",
            "flatten_4 (Flatten)          (None, 6272)              0         \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 6272)              0         \n",
            "_________________________________________________________________\n",
            "dense_7 (Dense)              (None, 512)               3211776   \n",
            "_________________________________________________________________\n",
            "dense_8 (Dense)              (None, 1)                 513       \n",
            "=================================================================\n",
            "Total params: 3,453,121\n",
            "Trainable params: 3,453,121\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AkQCvVhTLERJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "362c4f72-c77f-4ad9-d0ca-1da23f28b2bd"
      },
      "source": [
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1./255,\n",
        "    rotation_range=40,\n",
        "    width_shift_range=0.2,\n",
        "    height_shift_range=0.2,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True,)\n",
        "\n",
        "# Note that the validation data should not be augmented!\n",
        "test_datagen = ImageDataGenerator(rescale=1./255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "        # This is the target directory\n",
        "        train_dir,\n",
        "        # All images will be resized to 150x150\n",
        "        target_size=(150, 150),\n",
        "        batch_size=32,\n",
        "        # Since we use binary_crossentropy loss, we need binary labels\n",
        "        class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "        validation_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=32,\n",
        "        class_mode='binary')\n",
        "\n",
        "history = model.fit_generator(\n",
        "      train_generator,\n",
        "      steps_per_epoch=100,\n",
        "      epochs=30,\n",
        "      validation_data=validation_generator,\n",
        "      validation_steps=50)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 148 images belonging to 2 classes.\n",
            "Found 32 images belonging to 2 classes.\n",
            "Epoch 1/30\n",
            "100/100 [==============================] - 29s 294ms/step - loss: 0.3713 - acc: 0.8358 - val_loss: 0.7610 - val_acc: 0.7188\n",
            "Epoch 2/30\n",
            "100/100 [==============================] - 29s 288ms/step - loss: 0.3814 - acc: 0.8358 - val_loss: 0.4863 - val_acc: 0.7812\n",
            "Epoch 3/30\n",
            "100/100 [==============================] - 29s 289ms/step - loss: 0.3451 - acc: 0.8551 - val_loss: 0.7489 - val_acc: 0.7188\n",
            "Epoch 4/30\n",
            "100/100 [==============================] - 29s 288ms/step - loss: 0.3187 - acc: 0.8733 - val_loss: 0.4863 - val_acc: 0.7500\n",
            "Epoch 5/30\n",
            "100/100 [==============================] - 28s 285ms/step - loss: 0.3149 - acc: 0.8706 - val_loss: 0.5379 - val_acc: 0.7812\n",
            "Epoch 6/30\n",
            "100/100 [==============================] - 29s 294ms/step - loss: 0.2977 - acc: 0.8682 - val_loss: 0.5615 - val_acc: 0.7500\n",
            "Epoch 7/30\n",
            "100/100 [==============================] - 32s 317ms/step - loss: 0.2857 - acc: 0.8804 - val_loss: 0.5290 - val_acc: 0.8125\n",
            "Epoch 8/30\n",
            "100/100 [==============================] - 32s 316ms/step - loss: 0.2584 - acc: 0.8939 - val_loss: 0.5767 - val_acc: 0.7500\n",
            "Epoch 9/30\n",
            "100/100 [==============================] - 30s 298ms/step - loss: 0.2705 - acc: 0.8865 - val_loss: 0.5613 - val_acc: 0.8438\n",
            "Epoch 10/30\n",
            "100/100 [==============================] - 30s 295ms/step - loss: 0.2495 - acc: 0.8892 - val_loss: 0.7024 - val_acc: 0.8438\n",
            "Epoch 11/30\n",
            "100/100 [==============================] - 29s 294ms/step - loss: 0.2496 - acc: 0.9044 - val_loss: 0.5985 - val_acc: 0.7500\n",
            "Epoch 12/30\n",
            "100/100 [==============================] - 29s 293ms/step - loss: 0.2368 - acc: 0.9030 - val_loss: 0.6247 - val_acc: 0.7500\n",
            "Epoch 13/30\n",
            "100/100 [==============================] - 29s 289ms/step - loss: 0.2201 - acc: 0.9152 - val_loss: 0.5318 - val_acc: 0.7812\n",
            "Epoch 14/30\n",
            "100/100 [==============================] - 30s 298ms/step - loss: 0.2141 - acc: 0.9196 - val_loss: 0.5956 - val_acc: 0.8438\n",
            "Epoch 15/30\n",
            "100/100 [==============================] - 31s 306ms/step - loss: 0.2066 - acc: 0.9216 - val_loss: 0.5492 - val_acc: 0.8125\n",
            "Epoch 16/30\n",
            "100/100 [==============================] - 29s 293ms/step - loss: 0.2004 - acc: 0.9209 - val_loss: 0.7594 - val_acc: 0.7812\n",
            "Epoch 17/30\n",
            "100/100 [==============================] - 29s 292ms/step - loss: 0.1930 - acc: 0.9199 - val_loss: 0.8496 - val_acc: 0.7812\n",
            "Epoch 18/30\n",
            "100/100 [==============================] - 30s 297ms/step - loss: 0.1779 - acc: 0.9321 - val_loss: 0.7042 - val_acc: 0.7812\n",
            "Epoch 19/30\n",
            "100/100 [==============================] - 29s 290ms/step - loss: 0.1734 - acc: 0.9304 - val_loss: 0.7661 - val_acc: 0.8125\n",
            "Epoch 20/30\n",
            "100/100 [==============================] - 30s 295ms/step - loss: 0.1567 - acc: 0.9409 - val_loss: 1.0961 - val_acc: 0.6250\n",
            "Epoch 21/30\n",
            "100/100 [==============================] - 31s 310ms/step - loss: 0.1519 - acc: 0.9405 - val_loss: 0.6280 - val_acc: 0.8125\n",
            "Epoch 22/30\n",
            "100/100 [==============================] - 31s 308ms/step - loss: 0.1540 - acc: 0.9416 - val_loss: 1.3521 - val_acc: 0.7188\n",
            "Epoch 23/30\n",
            "100/100 [==============================] - 31s 306ms/step - loss: 0.1329 - acc: 0.9507 - val_loss: 0.6915 - val_acc: 0.8125\n",
            "Epoch 24/30\n",
            "100/100 [==============================] - 30s 303ms/step - loss: 0.1341 - acc: 0.9503 - val_loss: 0.7928 - val_acc: 0.8125\n",
            "Epoch 25/30\n",
            "100/100 [==============================] - 29s 289ms/step - loss: 0.1226 - acc: 0.9497 - val_loss: 0.7721 - val_acc: 0.8125\n",
            "Epoch 26/30\n",
            "100/100 [==============================] - 29s 292ms/step - loss: 0.1315 - acc: 0.9514 - val_loss: 0.8322 - val_acc: 0.7500\n",
            "Epoch 27/30\n",
            "100/100 [==============================] - 29s 291ms/step - loss: 0.1185 - acc: 0.9530 - val_loss: 0.8317 - val_acc: 0.8125\n",
            "Epoch 28/30\n",
            "100/100 [==============================] - 29s 288ms/step - loss: 0.1177 - acc: 0.9497 - val_loss: 0.7934 - val_acc: 0.8125\n",
            "Epoch 29/30\n",
            "100/100 [==============================] - 29s 291ms/step - loss: 0.1081 - acc: 0.9581 - val_loss: 1.1553 - val_acc: 0.7188\n",
            "Epoch 30/30\n",
            "100/100 [==============================] - 29s 289ms/step - loss: 0.1097 - acc: 0.9551 - val_loss: 1.0025 - val_acc: 0.8125\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dgf8G7uxqiK1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "model.save('clean_and_messy_aug.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fJ6cAz47SnqM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "outputId": "6367d603-37ca-4e59-83c8-f3eb0909832b"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "acc = history.history['acc']\n",
        "val_acc = history.history['val_acc']\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "epochs = range(len(acc))\n",
        "\n",
        "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.legend()\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lbz3RBpgLQac",
        "colab_type": "text"
      },
      "source": [
        "---"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A_bRqA51Ga5T",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# remove 2 layers\n",
        "smaller_model = models.Sequential()\n",
        "smaller_model.add(layers.Conv2D(32, (3, 3), activation='relu',\n",
        "                        input_shape=(150, 150, 3)))\n",
        "smaller_model.add(layers.MaxPooling2D((2, 2)))\n",
        "smaller_model.add(layers.Conv2D(32, (3, 3), activation='relu'))\n",
        "smaller_model.add(layers.MaxPooling2D((2, 2)))\n",
        "smaller_model.add(layers.Conv2D(32, (3, 3), activation='relu'))\n",
        "smaller_model.add(layers.MaxPooling2D((2, 2)))\n",
        "#smaller_model.add(layers.Conv2D(128, (3, 3), activation='relu'))\n",
        "#smaller_model.add(layers.MaxPooling2D((2, 2)))\n",
        "smaller_model.add(layers.Flatten())\n",
        "smaller_model.add(layers.Dropout(0.5))\n",
        "smaller_model.add(layers.Dense(64, activation='relu'))\n",
        "smaller_model.add(layers.Dense(1, activation='sigmoid'))\n",
        "\n",
        "smaller_model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.RMSprop(lr=1e-4),\n",
        "              metrics=['acc'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AtmicGOfRvmk",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 510
        },
        "outputId": "2071e52d-97aa-4165-d4b9-800ccd1d4f8c"
      },
      "source": [
        "smaller_model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_11\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d_31 (Conv2D)           (None, 148, 148, 32)      896       \n",
            "_________________________________________________________________\n",
            "max_pooling2d_31 (MaxPooling (None, 74, 74, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_32 (Conv2D)           (None, 72, 72, 32)        9248      \n",
            "_________________________________________________________________\n",
            "max_pooling2d_32 (MaxPooling (None, 36, 36, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_33 (Conv2D)           (None, 34, 34, 32)        9248      \n",
            "_________________________________________________________________\n",
            "max_pooling2d_33 (MaxPooling (None, 17, 17, 32)        0         \n",
            "_________________________________________________________________\n",
            "flatten_11 (Flatten)         (None, 9248)              0         \n",
            "_________________________________________________________________\n",
            "dropout_8 (Dropout)          (None, 9248)              0         \n",
            "_________________________________________________________________\n",
            "dense_21 (Dense)             (None, 64)                591936    \n",
            "_________________________________________________________________\n",
            "dense_22 (Dense)             (None, 1)                 65        \n",
            "=================================================================\n",
            "Total params: 611,393\n",
            "Trainable params: 611,393\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "34L8LgtHG7i8",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "outputId": "b70d86c9-e010-469f-8b51-9dffc7e3519e"
      },
      "source": [
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1./255,\n",
        "    rotation_range=40,\n",
        "    width_shift_range=0.2,\n",
        "    height_shift_range=0.2,\n",
        "    shear_range=0.2,\n",
        "    zoom_range=0.2,\n",
        "    horizontal_flip=True,)\n",
        "\n",
        "# Note that the validation data should not be augmented!\n",
        "test_datagen = ImageDataGenerator(rescale=1./255)\n",
        "\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "        # This is the target directory\n",
        "        train_dir,\n",
        "        # All images will be resized to 150x150\n",
        "        target_size=(150, 150),\n",
        "        batch_size=32,\n",
        "        # Since we use binary_crossentropy loss, we need binary labels\n",
        "        class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "        validation_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=32,\n",
        "        class_mode='binary')\n",
        "\n",
        "history = smaller_model.fit_generator(\n",
        "      train_generator,\n",
        "      steps_per_epoch=100,\n",
        "      epochs=30,\n",
        "      validation_data=validation_generator,\n",
        "      validation_steps=50)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 148 images belonging to 2 classes.\n",
            "Found 32 images belonging to 2 classes.\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-25-29f6970b8264>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[1;32m     26\u001b[0m         class_mode='binary')\n\u001b[1;32m     27\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m---> 28\u001b[0;31m history = smaller_model.fit_generator(\n\u001b[0m\u001b[1;32m     29\u001b[0m       \u001b[0mtrain_generator\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m     30\u001b[0m       \u001b[0msteps_per_epoch\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;36m100\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'smaller_model' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Cmr06nWXb1lk",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "smaller_model.save('clean_and_messy_smaller.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZiLD7BAlIcZX",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "outputId": "b28a0472-0c8c-45a7-ac77-abb29a53b0d1"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "\n",
        "acc = history.history['acc']\n",
        "val_acc = history.history['val_acc']\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "epochs = range(len(acc))\n",
        "\n",
        "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.legend()\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9TWj2S5PYBal",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "52c241a4-4791-4dd7-9c3a-a14a384d545f"
      },
      "source": [
        "test_generator = test_datagen.flow_from_directory(\n",
        "        test_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=20,\n",
        "        class_mode='binary')\n",
        "\n",
        "test_loss, test_acc = smaller_model.evaluate_generator(test_generator, steps=50)\n",
        "print('test acc:', test_acc)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 32 images belonging to 2 classes.\n",
            "test acc: 0.78125\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TC6qu9BXZooJ",
        "colab_type": "text"
      },
      "source": [
        "---"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CTgSIOTbZnxt",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras.applications import VGG16\n",
        "\n",
        "conv_base = VGG16(weights='imagenet',\n",
        "                  include_top=False,\n",
        "                  input_shape=(150, 150, 3))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "t6vQziIIZ_eN",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 827
        },
        "outputId": "1d77da5b-4204-4dd5-9a56-c46820dd6d05"
      },
      "source": [
        "conv_base.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"vgg16\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_1 (InputLayer)         (None, 150, 150, 3)       0         \n",
            "_________________________________________________________________\n",
            "block1_conv1 (Conv2D)        (None, 150, 150, 64)      1792      \n",
            "_________________________________________________________________\n",
            "block1_conv2 (Conv2D)        (None, 150, 150, 64)      36928     \n",
            "_________________________________________________________________\n",
            "block1_pool (MaxPooling2D)   (None, 75, 75, 64)        0         \n",
            "_________________________________________________________________\n",
            "block2_conv1 (Conv2D)        (None, 75, 75, 128)       73856     \n",
            "_________________________________________________________________\n",
            "block2_conv2 (Conv2D)        (None, 75, 75, 128)       147584    \n",
            "_________________________________________________________________\n",
            "block2_pool (MaxPooling2D)   (None, 37, 37, 128)       0         \n",
            "_________________________________________________________________\n",
            "block3_conv1 (Conv2D)        (None, 37, 37, 256)       295168    \n",
            "_________________________________________________________________\n",
            "block3_conv2 (Conv2D)        (None, 37, 37, 256)       590080    \n",
            "_________________________________________________________________\n",
            "block3_conv3 (Conv2D)        (None, 37, 37, 256)       590080    \n",
            "_________________________________________________________________\n",
            "block3_pool (MaxPooling2D)   (None, 18, 18, 256)       0         \n",
            "_________________________________________________________________\n",
            "block4_conv1 (Conv2D)        (None, 18, 18, 512)       1180160   \n",
            "_________________________________________________________________\n",
            "block4_conv2 (Conv2D)        (None, 18, 18, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "block4_conv3 (Conv2D)        (None, 18, 18, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "block4_pool (MaxPooling2D)   (None, 9, 9, 512)         0         \n",
            "_________________________________________________________________\n",
            "block5_conv1 (Conv2D)        (None, 9, 9, 512)         2359808   \n",
            "_________________________________________________________________\n",
            "block5_conv2 (Conv2D)        (None, 9, 9, 512)         2359808   \n",
            "_________________________________________________________________\n",
            "block5_conv3 (Conv2D)        (None, 9, 9, 512)         2359808   \n",
            "_________________________________________________________________\n",
            "block5_pool (MaxPooling2D)   (None, 4, 4, 512)         0         \n",
            "=================================================================\n",
            "Total params: 14,714,688\n",
            "Trainable params: 14,714,688\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2jZKPheoaxkh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras import models\n",
        "from keras import layers\n",
        "\n",
        "VGG16_model = models.Sequential()\n",
        "VGG16_model.add(conv_base)\n",
        "VGG16_model.add(layers.Flatten())\n",
        "VGG16_model.add(layers.Dense(256, activation='relu'))\n",
        "VGG16_model.add(layers.Dense(1, activation='sigmoid'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KdAJ_jXRa5aO",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 299
        },
        "outputId": "fd771f1e-d96b-4678-bd42-dc812dd0119b"
      },
      "source": [
        "VGG16_model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "vgg16 (Model)                (None, 4, 4, 512)         14714688  \n",
            "_________________________________________________________________\n",
            "flatten_1 (Flatten)          (None, 8192)              0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 256)               2097408   \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 1)                 257       \n",
            "=================================================================\n",
            "Total params: 16,812,353\n",
            "Trainable params: 16,812,353\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fhdCrKZRbOOY",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "f78cfa17-d73d-4f5e-a157-c714f62f05f2"
      },
      "source": [
        "print('This is the number of trainable weights '\n",
        "      'before freezing the conv base:', len(VGG16_model.trainable_weights))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "This is the number of trainable weights before freezing the conv base: 30\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ioBKQBd6bJWW",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "conv_base.trainable = False"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3rCMmjO7bNgB",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "outputId": "88337584-f1fc-49c0-c5da-bbb028dd0b2c"
      },
      "source": [
        "print('This is the number of trainable weights '\n",
        "      'after freezing the conv base:', len(VGG16_model.trainable_weights))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "This is the number of trainable weights after freezing the conv base: 4\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "n0JtuFJLbh2L",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "3cd915b1-adf8-4318-f5fe-9b9f3c1d3682"
      },
      "source": [
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras import optimizers\n",
        "train_datagen = ImageDataGenerator(\n",
        "      rescale=1./255,\n",
        "      rotation_range=40,\n",
        "      width_shift_range=0.2,\n",
        "      height_shift_range=0.2,\n",
        "      shear_range=0.2,\n",
        "      zoom_range=0.2,\n",
        "      horizontal_flip=True,\n",
        "      fill_mode='nearest')\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1./255)\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "        # This is the target directory\n",
        "        train_dir,\n",
        "        # All images will be resized to 150x150\n",
        "        target_size=(150, 150),\n",
        "        batch_size=20,\n",
        "        # Since we use binary_crossentropy loss, we need binary labels\n",
        "        class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "        validation_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=20,\n",
        "        class_mode='binary')\n",
        "\n",
        "VGG16_model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.RMSprop(lr=2e-5),\n",
        "              metrics=['acc'])\n",
        "\n",
        "history = VGG16_model.fit_generator(\n",
        "      train_generator,\n",
        "      steps_per_epoch=100,\n",
        "      epochs=30,\n",
        "      validation_data=validation_generator,\n",
        "      validation_steps=50,\n",
        "      verbose=2)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 148 images belonging to 2 classes.\n",
            "Found 32 images belonging to 2 classes.\n",
            "Epoch 1/30\n",
            " - 27s - loss: 0.4065 - acc: 0.8669 - val_loss: 0.2892 - val_acc: 0.8750\n",
            "Epoch 2/30\n",
            " - 20s - loss: 0.2265 - acc: 0.9279 - val_loss: 0.2334 - val_acc: 0.8750\n",
            "Epoch 3/30\n",
            " - 19s - loss: 0.1710 - acc: 0.9431 - val_loss: 0.2326 - val_acc: 0.8750\n",
            "Epoch 4/30\n",
            " - 20s - loss: 0.1429 - acc: 0.9580 - val_loss: 0.4389 - val_acc: 0.9062\n",
            "Epoch 5/30\n",
            " - 19s - loss: 0.1200 - acc: 0.9666 - val_loss: 0.2673 - val_acc: 0.8750\n",
            "Epoch 6/30\n",
            " - 19s - loss: 0.1012 - acc: 0.9729 - val_loss: 0.1376 - val_acc: 0.8750\n",
            "Epoch 7/30\n",
            " - 19s - loss: 0.0950 - acc: 0.9709 - val_loss: 0.1873 - val_acc: 0.8750\n",
            "Epoch 8/30\n",
            " - 19s - loss: 0.0914 - acc: 0.9680 - val_loss: 0.2077 - val_acc: 0.8750\n",
            "Epoch 9/30\n",
            " - 19s - loss: 0.0814 - acc: 0.9778 - val_loss: 0.4112 - val_acc: 0.8750\n",
            "Epoch 10/30\n",
            " - 19s - loss: 0.0744 - acc: 0.9795 - val_loss: 0.2505 - val_acc: 0.9375\n",
            "Epoch 11/30\n",
            " - 19s - loss: 0.0790 - acc: 0.9741 - val_loss: 0.0971 - val_acc: 0.9062\n",
            "Epoch 12/30\n",
            " - 19s - loss: 0.0673 - acc: 0.9821 - val_loss: 0.1815 - val_acc: 0.9062\n",
            "Epoch 13/30\n",
            " - 19s - loss: 0.0652 - acc: 0.9795 - val_loss: 0.0857 - val_acc: 0.9062\n",
            "Epoch 14/30\n",
            " - 19s - loss: 0.0531 - acc: 0.9892 - val_loss: 0.0120 - val_acc: 0.9062\n",
            "Epoch 15/30\n",
            " - 18s - loss: 0.0539 - acc: 0.9849 - val_loss: 0.0972 - val_acc: 0.9062\n",
            "Epoch 16/30\n",
            " - 19s - loss: 0.0493 - acc: 0.9886 - val_loss: 0.2950 - val_acc: 0.9062\n",
            "Epoch 17/30\n",
            " - 18s - loss: 0.0436 - acc: 0.9887 - val_loss: 0.0161 - val_acc: 0.9062\n",
            "Epoch 18/30\n",
            " - 19s - loss: 0.0450 - acc: 0.9864 - val_loss: 0.1123 - val_acc: 0.9062\n",
            "Epoch 19/30\n",
            " - 18s - loss: 0.0367 - acc: 0.9919 - val_loss: 0.1799 - val_acc: 0.8750\n",
            "Epoch 20/30\n",
            " - 19s - loss: 0.0423 - acc: 0.9876 - val_loss: 0.0067 - val_acc: 0.9062\n",
            "Epoch 21/30\n",
            " - 17s - loss: 0.0414 - acc: 0.9897 - val_loss: 0.0981 - val_acc: 0.9062\n",
            "Epoch 22/30\n",
            " - 19s - loss: 0.0357 - acc: 0.9892 - val_loss: 0.1644 - val_acc: 0.9062\n",
            "Epoch 23/30\n",
            " - 18s - loss: 0.0297 - acc: 0.9940 - val_loss: 0.3039 - val_acc: 0.9062\n",
            "Epoch 24/30\n",
            " - 18s - loss: 0.0362 - acc: 0.9892 - val_loss: 0.2272 - val_acc: 0.9062\n",
            "Epoch 25/30\n",
            " - 18s - loss: 0.0376 - acc: 0.9876 - val_loss: 0.1817 - val_acc: 0.9062\n",
            "Epoch 26/30\n",
            " - 19s - loss: 0.0317 - acc: 0.9913 - val_loss: 0.2554 - val_acc: 0.9062\n",
            "Epoch 27/30\n",
            " - 18s - loss: 0.0271 - acc: 0.9946 - val_loss: 0.0983 - val_acc: 0.9375\n",
            "Epoch 28/30\n",
            " - 18s - loss: 0.0321 - acc: 0.9919 - val_loss: 0.0809 - val_acc: 0.9375\n",
            "Epoch 29/30\n",
            " - 19s - loss: 0.0288 - acc: 0.9924 - val_loss: 0.1661 - val_acc: 0.9062\n",
            "Epoch 30/30\n",
            " - 19s - loss: 0.0249 - acc: 0.9946 - val_loss: 0.1484 - val_acc: 0.8750\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R8Ju66jucN9h",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "VGG16_model.save('clean_and_messy_VGG16.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JocgdOpnboH2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "outputId": "835a0e58-9ef4-4f4e-dde4-95f0885ffc5c"
      },
      "source": [
        "acc = history.history['acc']\n",
        "val_acc = history.history['val_acc']\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "epochs = range(len(acc))\n",
        "\n",
        "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.legend()\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXsAAAEICAYAAAC+iFRkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO29ebgU1bnv/3k3owjI6MQWEAURw7wBsbeK0eTg8JML0QQkicQkRM3ojckxV6Necvjl5EiMJ09icvHEmYQYb2LwBmOM0Wt2i8oGAWVSRGRSRCZBQKb3/rGqNr2b7t1V3dXz+3mefrp61apVb3V1f2utd631LlFVDMMwjMqmptgGGIZhGPnHxN4wDKMKMLE3DMOoAkzsDcMwqgATe8MwjCrAxN4wDKMKMLGvQkTkKRG5Nuq8xURE1onIJXkoV0XkTG/71yLywyB5szjPVBH5W7Z2GkYmxMbZlwcisifhYwfgY+Cw9/lrqjqn8FaVDiKyDviKqv494nIV6K+qa6LKKyJ9gbeBNqp6KAo7DSMTrYttgBEMVe3ob7ckbCLS2gTEKBXs91g6mBunzBGRcSKyUUT+VUTeAx4Qka4i8n9EZKuI7PC2axOOeV5EvuJtTxORBhGZ5eV9W0QuzTLv6SLygojsFpG/i8gvReTRNHYHsfFHIhL3yvubiPRI2P8FEXlHRLaJyK0tfD9jROQ9EWmVkDZRRJZ526NFZIGI7BSRd0XkFyLSNk1ZD4rIvyV8/p53zGYRuS4p7+Ui8qqIfCgiG0TkzoTdL3jvO0Vkj4iM9b/bhOPPE5GFIrLLez8v6HcT8nvuJiIPeNewQ0SeSNg3QUSWeNfwloiM99KbucxE5E7/PotIX8+d9WURWQ/8w0v/g3cfdnm/kXMSjj9ORH7q3c9d3m/sOBH5i4h8M+l6lonIxFTXarSMiX1lcDLQDegDTMfd1we8z72BfcAvWjh+DLAa6AH8B/AbEZEs8v4WeAXoDtwJfKGFcwax8RrgS8CJQFvgZgARGQT8yiv/VO98taRAVV8GPgI+mVTub73tw8BN3vWMBS4GbmzBbjwbxnv2fAroDyT3F3wEfBHoAlwO3CAi/83bd4H33kVVO6rqgqSyuwF/AX7uXdvdwF9EpHvSNRzz3aQg0/f8CM4teI5X1s88G0YDDwPf867hAmBduu8jBRcCZwP/4n1+Cvc9nQgsBhLdjrOAkcB5uN/x94EjwEPA5/1MIjIU6IX7boywqKq9yuyF+9Nd4m2PAw4A7VvIPwzYkfD5eZwbCGAasCZhXwdAgZPD5MUJySGgQ8L+R4FHA15TKhtvS/h8I/BXb/t2YG7CvuO97+CSNGX/G3C/t90JJ8R90uT9DvCnhM8KnOltPwj8m7d9P/DvCfkGJOZNUe49wM+87b5e3tYJ+6cBDd72F4BXko5fAEzL9N2E+Z6BU3Ci2jVFvv/l29vS78/7fKd/nxOurV8LNnTx8pyAexjtA4amyNce2IHrBwH3ULi30P+3SnlZzb4y2Kqq+/0PItJBRP6X1yz+EOc26JLoykjiPX9DVfd6mx1D5j0V2J6QBrAhncEBbXwvYXtvgk2nJpatqh8B29KdC1eLnyQi7YBJwGJVfcezY4Dn2njPs+P/x9XyM9HMBuCdpOsbIyLPee6TXcD1Acv1y34nKe0dXK3WJ91304wM3/NpuHu2I8WhpwFvBbQ3FU3fjYi0EpF/91xBH3K0hdDDe7VPdS7vN/174PMiUgNMwbVEjCwwsa8MkodUfRc4Cxijqp056jZI55qJgneBbiLSISHttBby52Lju4lle+fsni6zqq7AieWlNHfhgHMHrcLVHjsD/yMbG3Atm0R+C8wDTlPVE4BfJ5SbaQjcZpzbJZHewKYAdiXT0ve8AXfPuqQ4bgNwRpoyP8K16nxOTpEn8RqvASbgXF0n4Gr/vg0fAPtbONdDwFSce22vJrm8jOCY2FcmnXBN452e//eOfJ/Qqyk3AneKSFsRGQv8f3my8XHgChGp9zpTZ5D5t/xb4Ns4sftDkh0fAntEZCBwQ0AbHgOmicgg72GTbH8nXK15v+f/viZh31ac+6RfmrLnAwNE5BoRaS0inwMGAf8noG3JdqT8nlX1XZwv/V6vI7eNiPgPg98AXxKRi0WkRkR6ed8PwBJgspe/DrgqgA0f41pfHXCtJ9+GIziX2N0icqrXChjrtcLwxP0I8FOsVp8TJvaVyT3Acbha00vAXwt03qm4Ts5tOD/573F/8lRkbaOqLge+jhPwd3F+3Y0ZDvsdrtPwH6r6QUL6zTgh3g3c59kcxIanvGv4B7DGe0/kRmCGiOzG9TE8lnDsXmAmEBc3CujcpLK3AVfgauXbcB2WVyTZHZRM3/MXgIO41s37uD4LVPUVXAfwz4BdwP/laGvjh7ia+A7gf9K8pZSKh3Etq03ACs+ORG4GXgMWAtuBn9Bcmx4GBuP6gIwssUlVRt4Qkd8Dq1Q17y0Lo3IRkS8C01W1vti2lDNWszciQ0RGicgZXrN/PM5P+0Sm4wwjHZ6L7EZgdrFtKXdM7I0oORk3LHAPboz4Dar6alEtMsoWEfkXXP/GFjK7iowMZHTjiMj9OP/h+6r6iRT7BfhP4DLcELBpqrrY23ctcJuX9d9U9aEIbTcMwzACEqRm/yAwvoX9l+JmxvXHzd78FTTNArwDN+NyNHCHiHTNxVjDMAwjOzIGQlPVF8RF6UvHBOBhdU2El0Ski4icgpvZ+YyqbgcQkWdwD43ftXS+Hj16aN++LZ3OMAzDSGbRokUfqGrPdPujiHrZi+YzCTd6aenSj0FEpuNaBfTu3ZvGxsYIzDIMw6geRCR51nUzSqKDVlVnq2qdqtb17Jn2wWQYhmFkSRRiv4nm08ZrvbR06YZhGEaBiULs5wFfFMe5wC5vGvbTwKe9adhdgU97aYZhGEaByeizF5Hf4Tpbe4jIRtwImzYAqvprXByPy3BTxvfiplijqttF5Ee4KdAAM/zO2rAcPHiQjRs3sn///syZjaLQvn17amtradOmTbFNMQwjBSUXLqGurk6TO2jffvttOnXqRPfu3Um/poZRLFSVbdu2sXv3bk4//fRim2MYVYmILFLVunT7S6KDNhP79+83oS9hRITu3btby8soGHPmQN++UFPj3ufMyXSEUTYLjpvQlzZ2f4xCMWcOTJ8Oe71lct55x30GmDq1eHaVOmVRszcMw/C59dajQu+zd69LN9JjYh+Abdu2MWzYMIYNG8bJJ59Mr169mj4fOHCgxWMbGxv51re+lfEc5513XlTmGkZZEtQ1s3598PR8uHvKpcxjKPYiuMmvkSNHajIrVqw4Jq0lHn1UtU8fVRH3/uijoQ5vkTvuuEPvuuuuZmkHDx6M7gRlTNj7ZBg+jz6q2qGDKhx9deiQ+r/bp0/zfP6rT5/sy8yHnYUuE2jUalpw3PfnvfOO+9p8f17UT8pp06Zx/fXXM2bMGL7//e/zyiuvMHbsWIYPH855553H6tWrAXj++ee54oorALjzzju57rrrGDduHP369ePnP/95U3kdO3Zsyj9u3DiuuuoqBg4cyNSpU1FvxNT8+fMZOHAgI0eO5Fvf+lZTuYmsW7eO888/nxEjRjBixAhefPHFpn0/+clPGDx4MEOHDuWWW24BYM2aNVxyySUMHTqUESNG8NZbuawxbVQLUddEw7hmZs6EDh2ap3Xo4NKzLTMfdhazzJS09CQoxivXmn3Qp362+DX7a6+9Vi+//HI9dOiQqqru2rWrqYb/zDPP6KRJk1RV9bnnntPLL7+86dixY8fq/v37devWrdqtWzc9cOCAqqoef/zxTfk7d+6sGzZs0MOHD+u5556r//znP3Xfvn1aW1ura9euVVXVyZMnN5WbyEcffaT79u1TVdU33nhD/e9z/vz5OnbsWP3oo49UVXXbtm2qqjp69Gj94x//qKqq+/bta9qfDVazrw7yUbsVSf2/FUlvQ6bWe9gyy91Oqq1mH8aflytXX301rVq1AmDXrl1cffXVfOITn+Cmm25i+fLlKY+5/PLLadeuHT169ODEE09ky5Ytx+QZPXo0tbW11NTUMGzYMNatW8eqVavo169f0zj2KVOmpCz/4MGDfPWrX2Xw4MFcffXVrFixAoC///3vfOlLX6KDVyXq1q0bu3fvZtOmTUycOBFwE6M6JFeZDCOJsDXRIK2A3r1TH5sufepUWLcOjhxx76lG4YQtM2o7g3oZwtqZLRUn9oX64gCOP/74pu0f/vCHXHTRRbz++us8+eSTacect2vXrmm7VatWHDp0KKs86fjZz37GSSedxNKlS2lsbMzYgWyUL8Uaax62gzSI4AV1zYQhTJn5sDPoQzEf156KihP7Qn1xyezatYtevVwE5wcffDDy8s866yzWrl3LunXrAPj973+f1o5TTjmFmpoaHnnkEQ4fPgzApz71KR544AH2er++7du306lTJ2pra3niCbdM7Mcff9y03yhtwvRNRf1QCFOhCip4U6fC7NnQpw+IuPfZs3MbNx+mzHzYGfShmI9rT0XFiX2hvrhkvv/97/ODH/yA4cOHh6qJB+W4447j3nvvZfz48YwcOZJOnTpxwgknHJPvxhtv5KGHHmLo0KGsWrWqqfUxfvx4rrzySurq6hg2bBizZs0C4JFHHuHnP/85Q4YM4bzzzuO9996L3HYjuOAGzRdUnPIxYCFMhSpMKyCIayYsQcvMh51hHor5uPZjaMmhX4xXFEMvK5Xdu3erquqRI0f0hhtu0LvvvrvIFjXH7lNqgnZohun4DNqpl68BC0GHN+d7wERU5MPOfHRktwQZOmiLLu7JLxP79Nx99906dOhQPfvss/Waa67JaeRMPrD7lJqgQhJGcILmzceIlDAUWvCyJV925nPOTzIm9kbBsPuUmqCCG0aYg4pTKdSsCyl4uVAudqYjk9hXnM/eMApFUP96UN9tWB9vkL6psAMW8jHCpyD+6AgoFzuzpqUnQTFeVrMvX6rpPoVp9ufDZx/W1iA11nJxuRipwdw4RqGopvsU1j0SRnCL5UooBZePkT2ZxN7cOIaRRBBXRtiZ2kFdBMV0JRRy9rlReEzsA3DRRRfx9NPN10q/5557uOGGG9IeM27cOPzlFS+77DJ27tx5TJ4777yzabx7Op544ommkAcAt99+O3//+9/DmG+EoNSmuBeSSrwm4ygm9gGYMmUKc+fObZY2d+7ctPFpkpk/fz5dunTJ6tzJYj9jxgwuueSSrMoyMlNqU9wLSSVek3GUQGIvIuNFZLWIrBGRW1Ls7yMiz4rIMhF5XkRqE/b9h4gsF5GVIvJzKcP166666ir+8pe/NMWZWbduHZs3b+b888/nhhtuoK6ujnPOOYc77rgj5fF9+/blgw8+AGDmzJkMGDCA+vr6pjDIAPfddx+jRo1i6NChfOYzn2Hv3r28+OKLzJs3j+9973sMGzaMt956i2nTpvH4448D8OyzzzJ8+HAGDx7Mddddx8cff9x0vjvuuIMRI0YwePBgVq1adYxNFgo5NaU2xb2QVOI1GUfJuAatiLQCfgl8CtgILBSReaq6IiHbLOBhVX1IRD4J/Bj4goicB8SAIV6+BuBC4PlsDf7Od2DJkmyPTs2wYXDPPen3d+vWjdGjR/PUU08xYcIE5s6dy2c/+1lEhJkzZ9KtWzcOHz7MxRdfzLJlyxgyZEjKchYtWsTcuXNZsmQJhw4dYsSIEYwcORKASZMm8dWvfhWA2267jd/85jd885vf5Morr+SKK67gqquualbW/v37mTZtGs8++ywDBgzgi1/8Ir/61a/4zne+A0CPHj1YvHgx9957L7NmzeK//uu/mh1/4okn8swzz9C+fXvefPNNpkyZQmNjI0899RR//vOfefnll+nQoQPbt28HYOrUqdxyyy1MnDiR/fv3c+TIkay+61Knd2/nukmVnszUqZUnhJV4TYYjSM1+NLBGVdeq6gFgLjAhKc8g4B/e9nMJ+xVoD7QF2gFtgGNj+pYBia6cRBfOY489xogRIxg+fDjLly9v5nJJ5p///CcTJ06kQ4cOdO7cmSuvvLJp3+uvv87555/P4MGDmTNnTtoQyT6rV6/m9NNPZ8CAAQBce+21vPDCC037J02aBMDIkSObgqclUoxQyMWK0hgGc2UYlUrGmj3QC9iQ8HkjMCYpz1JgEvCfwESgk4h0V9UFIvIc8C4gwC9UdWXyCURkOjAdoHeG3qCWauD5ZMKECdx0000sXryYvXv3MnLkSN5++21mzZrFwoUL6dq1K9OmTUsb2jgT06ZN44knnmDo0KE8+OCDPP/88znZ64dJThciOTEU8pEjR2jfvn1O58uE3/Hp+8P9jk8orZqkb8uttzrXTe/eTuhLyUbDyIaoOmhvBi4UkVdxbppNwGERORM4G6jFPTQ+KSLnJx+sqrNVtU5V63r27BmRSdHSsWNHLrroIq677rqmWv2HH37I8ccfzwknnMCWLVt46qmnWizjggsu4IknnmDfvn3s3r2bJ598smnf7t27OeWUUzh48CBzEqq8nTp1Yvfu3ceUddZZZ7Fu3TrWrFkDuOiVF154YeDrKXQo5IItvRYBFT+T0qhKgoj9JuC0hM+1XloTqrpZVSep6nDgVi9tJ66W/5Kq7lHVPcBTwNhILC8CU6ZMYenSpU1iP3ToUIYPH87AgQO55ppriMViLR4/YsQIPve5zzF06FAuvfRSRo0a1bTvRz/6EWPGjCEWizFw4MCm9MmTJ3PXXXcxfPjwZp2i7du354EHHuDqq69m8ODB1NTUcP311we+lkKHQrYx3IZRXMRNvGohg0hr4A3gYpzILwSuUdXlCXl6ANtV9YiIzAQOq+rtIvI54KvAeJwb56/APar6ZPJ5fOrq6tQfn+6zcuVKzj777GyuzyggLd2nvn1Td3z26eNqz4Zh5IaILFLVunT7M9bsVfUQ8A3gaWAl8JiqLheRGSLi9zCOA1aLyBvASYDfnfU48BbwGs6vv7QloTcqF+v4NIziEqSDFlWdD8xPSrs9YftxnLAnH3cY+FqONhpFYNs22LQJDhyAtm2hVy/o3j378sJ0fM6ZYx2khhE1gcS+FFBVynA+VlmybZtzufhD6Q8cOOqCSSf4mdyBEGwMd7mM2jGMcqMswiW0b9+ebdu2BRIUI3c2bToq9D5Hjrj0VKgq27Zti2T4ZthRO1Gv7WoYlUpZ1Oxra2vZuHEjW7duLbYpVcHmzen3tWmTOr19+/bU1tam3hmCMKN2grYCrLVgGAFG4xSaVKNxjGgI6gsv5siZMOcOmtdGAhnVQM6jcYzKIGjoXgg3ciZq90iYcwdtBdgYf8OgPFaqMnInHysrFXsZvaDXZCswGdUAGVaqMjdOlVBT4yQuGZFjO2ODUmz3SLIvHlwrIDksb9B8hlHOmBvHAPKzClGx3SNB469bnHbDsA7aqiEftdti1+wNwzhKppp9WQy9NHInH6F7Z85M/QBJFwLh8GH48Y/BRtBm5vzzIWm9mpxoaIA//CFY3vbt4V//Fbp1i+bchw650ORf+xp06hRNmfng4EH4z/+E66+Hjh2LbU30WM3eyIkwoQ1efBFiMfdHam3VjLTs2+cemh984PpaoiAWg1deySxiqrBrF/ziF/D1r0dz7r/+FS69FO69F264IZoy88GTT8KVV8J998FXvlJsa8JjNXsjr4RZxi4ed+9r1sBJJ+XPpnLngQfguutg1SoYNCj38vbvh8ZGt6TnXXe1nFcVamvdvYpK7P37Ho+Xttgn2lmOYp8J66A1CkY8DmeeaUKfCX9ZBF98cmXRIhffKMNyC4DrwI7Fojs3NBfRUqZc7MwWE/sKoBzivqi6P1EQwal2+veHnj2jEx2/nPPOC5Y/FnNuuQ0bMufNxMGD8PLLzle/bl36+ErF5uOPYeFCZ+ebb8L77xfbougxsS9zwsyMLSZvvOF80PX1xbak9PFr1w0N0ZTX0AADBsCJJwbL79+jKB42S5a4Dnx/EbVSrTUvWuQEv9TtzAUT+xIlaG29XNZ29f88VrMPRiwGb70FW7bkVo7q0Y7xoAwdCscfH43g+WV8/euu07lURdS365vfhHbtStfOXDCxL0HC1NaLPbEpKPG4G8p31lnFtqQ8iMpvv3q1W58gjNi3bg1jxkQn9n37urkXo0eXrojG4859dtppMGpU6dqZCyb2JUiY2no+Zsbmg4YGJzhRDSWsdEaMiKaG6buCwrrP6uth6VLYvTv7c6seve9+mUuWwJ492ZeZD5L7k2Ix59bZt6+4dkWN/fVKkDC19XJY23XrVuezNxdOcNq1czXhXP328Tj06OF89mGIxVzMpJdeyv7cb78N773XXEQPH3YdtqWE35+UaOfBg67DtpIwsS9BwtTWyyHuy4svuncT+3DEYrB48bGtvDDE424UTtgVPc8917XCcmlZJPfTjB3r7Cg1F0mynf6opVKzM1cCib2IjBeR1SKyRkRuSbG/j4g8KyLLROR5EalN2NdbRP4mIitFZIWI9I3O/MokbG196lQ3rO3IEfdeSkIPrnbati3UpZ3bZ6QiFnOhBrKtYW7Z4oYRZvOQ7dwZBg/OrWXR0AAnnADnnOM+n3CCK7PURLShwa2tPHCg+9y9O5x9dunZmSsZxV5EWgG/BC4FBgFTRCR5Xt8s4GFVHQLMAH6csO9h4C5VPRsYDVTgCNZoKYfaehjicSf0ESxRW1X4NcxsBddvUWU73LW+3rlxDh3K7vh43NXmW7U6mhaLwYIFzp1TKqRq/fgTy7IN/12KBKnZjwbWqOpaVT0AzAUmJOUZBPzD237O3+89FFqr6jMAqrpHVXNolFYPpV5bD8r+/a6zy1w44enWzYVLyLaGGY873//IkdkdH4vBRx/BsmXhj92xA5YvP/a+x2Ku0/e117KzKWrS9SfFYrBzJ6xcWRy78kEQse8FJM6l2+ilJbIUmORtTwQ6iUh3YACwU0T+KCKvishdXkuhGSIyXUQaRaTRFhWvLBobg0/VN47FrwlnU8P0W1Tt2mV/br+csCxY4N6TWxVRh4LIlXStnygnlpUKUXXQ3gxcKCKvAhcCm4DDuEBr53v7RwH9gGnJB6vqbFWtU9W6nj17RmSSUQr4LoigU/WN5vg1zBUrwh23b59rUeUyY7l3bzfuPBs3UkODG68/enTz9D59oFev0hFRvz8pufVzxhluxnFUs5hLgSBivwk4LeFzrZfWhKpuVtVJqjocuNVL24lrBSzxXECHgCeAEZFYbpQF8bibSGXP8OzwxTqs6Cxc6IYP5tqi8n3XYSOhx+MwfPixAw2iDgWRK+n6k/IREK7YBBH7hUB/ETldRNoCk4F5iRlEpIeI+GX9ALg/4dguIuL/1T8JhKyjGOXKkSPhp+obzenXz0UJDSs6YYOfpSMWc8HLwszIPnDAxc5P16qIxVyQtSgCreXCvn3OzZjOzvp6WLvWzRWoBDKKvVcj/wbwNLASeExVl4vIDBG50ss2DlgtIm8AJwEzvWMP41w4z4rIa4AA90V+FUZJsmoVbN9uwc9yIdsaZkODGz7YvXtu58+mZbF4seuYT/eQLxV/eGNjy62fUutfyJVAPntVna+qA1T1DFX1hfx2VZ3nbT+uqv29PF9R1Y8Tjn1GVYeo6mBVneaN6DGqAAt+Fg319W426ubNwfJH2aIaPNiF/Q0jeJnu+5Ah0QVay4VMrZ/hw517p1RcTrliM2gLTDnEno+KeNz56vv3L7Yl5U3YGubKla5TNwqxb9XKzaYNK/b9+sHJJ6fe37q1K7PYIur3J/XokXp/27alHbwtLCb2BaRcYs9HhR9cKuxUfaM5w4fDcccFF51sg5+lo77ejYvfuTNzXj/4WaZzx2Ju/H4ugdZywW/9ZLKzvh5efdXNNyh3TOwLSLnEno+CLVvcWrPmwsmdNm3C1TDjcTds8Iwzojl/LOZEPEhQtDVr3ESlTPe9vj73QGu54PcnZbLTD1nxyiuFsSufmNgXkHKJPR8F5q+PFr+GGSQ8cNQtqjFjnDsnyMMm6H33A60Vy5UT1M6xY5vnL2dM7AtIucSejwJ/qv4Im1URCX544Ew1zHffdcMFo3zIduzoVq8KKvZdu7qRQC3RqZPrqC2WiAbtT+raFT7xCRN7IyTlEHs+KhoanOsh26n6RnOChgf290c93NUPinbwYMv5Ghrc6JYgi9TEYrkFWssFf1GVIK2fUgzelg0m9hEQdIRNpUWzTMfevW6stblwoqNLl2A1zHjcDRccPjza88dibhLSkiXp82zb5nzhQe97fX32gdZyYcsWt75vUDtjMdi1ywV2K2dM7HMk7AibSolm2RILF7ramol9tMRibgRJSzXMeNy1qNq2jf7cfvnpCBtS2S+z0H77sK2fUpkElism9jlSTSNsgmLBz/KDHx749ddT7//oI9eiyseM5V69XKu1JWFuaHAjh4IuUnPaae5VaBFtaHCtn6D9SX37wimnmNhXPdU0wiYo8biLw96tW7EtqSwy1a5fecXV+vPVosoUFC0ed9EjjzsueJn19dkFWsuFeBxGjQre+im14G3ZYmKfI9U0wiYIR464zixz4URP375w6qnpRcd/CPjDBaMmFnNBwd5++9h9+/c7913Y+55NoLVcyLY/KRZzLtpNmzLnLVVM7HOkmkbYBGHFCjfT0oKfRU+moGgNDa4Tt2vX/Jy/paBoixa5aJdh73uh/fZ+f1JYOyvBb29inyPVMsImKP6f1mr2+SEWc7XgjRubpx8+nP8W1TnnuEXDUwletiGVswm0lgv+7zNs62foUFeJM7GvcqphhE1Q4nEXf71fv2JbUpmkq2EuXw4ffphfsa+pcSKZTuz793dhGsLQqlX6MvNBtv1Jbdq4mcTl7Lc3sTciJR53gmTBz/LD0KEuPHCy6ORrMlUy9fXuwbJjx9E01aP3PRtiseCB1nIhaPCzdNTXw9KlwUJWlCIm9kZkbN7sOu/MhZM/Wrd2NczkmnBDgxse2Ldvfs/v31t/TD3A6tVuQlW29z1MoLVcWL7cTY7Kxc7Dh+Hll6O1q1CY2BuRYcHPCkMs5mqYieGBCxVOevRo98BJfNjket/DBFrLhVztPPfcYCErSqmjPjgAABk9SURBVBUTeyMy4nE3xjrqqfpGc/zwwH4Nc+NGNyywEA/ZDh3c/U0W++7d3UIg2dCxIwwbVhixP/nk7PuTTjjBBW8rV7+9ib0RGfG4q6W1aVNsSyqb5PDAhfLX+9TXuwlcB7wFRsMEFUuHHxQtU6C1XIii9ePbWY5B0UzsjUjYs8fFWzcXTv7p3NkNWfRFPh53Ne6hQwtz/ljMTaJavBjefx/efDP3+x4k0FouRNWf5IeseO21aOwqJIHEXkTGi8hqEVkjIrek2N9HRJ4VkWUi8ryI1Cbt7ywiG0XkF1EZbpQW+Z6qbzQnMTxwoVtUiWEb/I7aKETULzMfRNWfVKzgbVGQUexFpBXwS+BSYBAwRUQGJWWbBTysqkOAGcCPk/b/CHghd3ONUiUed83jfE3VN5pTX+9aUy++6GrDhZyxfPLJbsnDePzoIjVBg5+lww+0lk+xj6I/qXdvqK0tz07aIDX70cAaVV2rqgeAucCEpDyDgH94288l7heRkcBJwN9yN9coVfyp+l26FNuS6sCvYd59t+usLXSLyg8M1tDghD6KRWr8MvMRFK2hIZrWT6aQFaVMELHvBWxI+LzRS0tkKTDJ254IdBKR7iJSA/wUuLmlE4jIdBFpFJHGrVu3BrO8AARdlKTaKcRUfaM5fg3zz392AnTuuYU9fyzmFhZ/6aXo7ntLgdZyYc8e1/qJ0s4NG8ovsm1UHbQ3AxeKyKvAhcAm4DBwIzBfVTe2dLCqzlbVOlWt69mzZ0Qm5UbYRUmqmddfd51WFvyssPjf95AhblhgMc6dvB1FmVHXmv3+pFK3M9+0DpBnE3BawudaL60JVd2MV7MXkY7AZ1R1p4iMBc4XkRuBjkBbEdmjqsd08haSjz6CxsaWm4vf/W76RUnKOfbNnj3udfLJ0ZVpwc+KQywGc+cW53sfONBF19yxI7pFavxAaw0N8IUvRFMmuPKi7E8aPNjNDfjTn9ys5Sjp3Dn4oiphCSL2C4H+InI6TuQnA9ckZhCRHsB2VT0C/AC4H0BVpybkmQbUFVvoAW67De65J7tjy63plsxNN8Ezz7imclSzLeNxF2e9T59oyjOCcdFF7h5efHHhz11TA5/8pFvLtXv36MrMR1C0eNz1J0XV+mndGi64AP7wB/eKkjFj8hc2IqPYq+ohEfkG8DTQCrhfVZeLyAygUVXnAeOAH4uI4kbdfD0/5kbD6tVutt+vf50+z+TJbmHiZMp9UZJnn3UuqbVr3YiKKCjUVH2jOeecA2+8Ed19DMtvfnN0YlVUxGLwwx+6FkMUcfn9/qSoW+MPP5yfsfadO0dfpk+Qmj2qOh+Yn5R2e8L248DjGcp4EHgwtIV5YP161wwdNy59np/+1PnoE1055b4oiT+xBJxARyESfkfVd7+be1lGeM48s3jnzkc/ge8PX7AALrss9/Ly1Z/UvXvL+lGKVOUM2g0bMtfQ/UVJ/LjXvXqV/6Ikic3jqCaFWPAzI0r8QGtR/T6tP+koVSf2u3a5RR6CuGOmTnVuD4Cf/KS8hR6OTiy55JLo/KLxuIuvXqip+kZlkyrQWi5Yf9JRqk7sN3gzBk47reV8PoVeNi2f+NPqx41za8Vu3x5Nmeee62pjhhEFyYHWcsEW0zlK1Ym9P5omaEdrq1ZOzMoxFkYiiYHKEv2iubB7t4urbk1kI0oSA63lgt+fZL9Ph4l9AOrrXUdPvpdNyyeJgcpGjYrGL/rSS8WZqm9UNlEFRbP+pOZUndhv2OCELsykokItm5ZPEgOVdejgJm5E8WeqqSn8VH2jskkMtJYL1p/UnKoT+/Xr3ciaVq2CH1OoZdPySXKgsvp6WLgQPv44+zLjcTdVP59jg43qxA82lktQNOtPak5Vin3YiVH+smnl6rdPFagsV7/ooUMW/MzIH7GYWxhlzZrsjrf+pGOpOrHfsCH4SJxEYjG35mc+l03LF/7EkmSxh+xbK8uWuRhD9mcy8kGuv0/rTzqWqhL7w4fd4szZhDzI97Jp+STVGqUnnZSbX7TQ654a1cXZZ7twCbn8Pq0/qTlVJfZbtriaebY1eyhPv31DQ+qJJfX12ftFGxrc95jNd2kYmaipcdE0s3WdNjRYf1IyVSX2/oSqbGr2/rJp5ei3TxeozF+A4s03w5WnerRMw8gXsRisWgXbtoU77tChaBdVqRSqSuyzGWOfSBQjBApNSxNLsm2trF8PmzaZC8fIL/7vy1/UPCh+f5L9PptTlWKfreshX8um5ZOWfOv+AhRhxd6CSxmFoK7OrRkbtjVtv8/UVJXYb9gA7du7YZTZrCvr/3jKyZXT0sSSmprsFk+Ox128oMGDo7HRMFJx3HEwcmR2v0/rTzqWqhL7eNxNIlq/Prt1Zf1l08qpk9YPfpZuYonvF/3gg3BlnntuuIlphpENsZhbQjTo5D/rT0pPVYn9smXH+tv9dWWD0KpVfpZNyxf+xJKWfJf+nyKoX3TXLrdCj/lDjUJQX++EftGiYPmtPyk9VSX26UKmhllXNhaD5cvdsmmlTpCJJaNGQdu2wR9gCxa4B6bVnIxC4C9mHtR1av769FSN2O/fn35f2AiYEH6EQDEIMrGkfXvnFw36Z4rHXQtnzJhobDSMljjxROjfP3hlxPqT0lM1Yr9xo3tv27Z5eth1Zf1l08rBlRN0YonvF23pgegTj7sO7o4do7HRMDIRZvJfPO5crdafdCyBxF5ExovIahFZIyK3pNjfR0SeFZFlIvK8iNR66cNEZIGILPf2fS7qCwiK76q5+WY3k1TEvYddVzbqZdPyRZiJJbGYc3Fl8osePGiTVYzCE4u5iVWrV7ecb+dO159kv8/UZBR7EWkF/BK4FBgETBGRQUnZZgEPq+oQYAbwYy99L/BFVT0HGA/cIyJdojI+DL7YX3cdrFvnfNnr1mW3rmwsFt2yafkiTKCyoJOrlixx8YHsz2QUkqC/z5desv6klghSsx8NrFHVtap6AJgLTEjKMwj4h7f9nL9fVd9Q1Te97c3A+0DPKAwPix8qobY297Lq66NZNi2fhFmlp2dPGDAgs9/eVv4xisFZZ0H37pnF3vqTWiaI2PcCNiR83uilJbIUmORtTwQ6iUj3xAwiMhpoC7yVfAIRmS4ijSLSuHXr1qC2h2L9ehfpsV273Msqh6BofqCyoJ3PsZjrdG7JL9rQ4Cai9Uq++4aRR0Tc7zNTZaShwfqTWiKqDtqbgQtF5FXgQmATcNjfKSKnAI8AX1LVI8kHq+psVa1T1bqePfNT8V+/ProZdSefDP36la7YZzOxJJNf1CarGMUkFnMB+95/P/X+gwfdehP2+0xPELHfBCTKZK2X1oSqblbVSao6HLjVS9sJICKdgb8At6pq0VZx3bAh+wBoqaivdzWJUgyK5k8sCfPD94eUpnuAvf22iwtkfyajGGSa/Gf9SZkJIvYLgf4icrqItAUmA/MSM4hIDxHxy/oBcL+X3hb4E67z9vHozA6HanbLEbaEHx4422XT8onf3A0zi3DAAOjRI31T2RYrMYpJXZ1zwaarjFh/UmYyir2qHgK+ATwNrAQeU9XlIjJDRK70so0DVovIG8BJgD9y/bPABcA0EVnivYZFfRGZ2LHDjUyJMjBSKfvts5lYIuJmK6a7noYGFxfonHOisdEwwtCunRP8dJUR60/KTCCfvarOV9UBqnqGqs700m5X1Xne9uOq2t/L8xVV/dhLf1RV26jqsIRXwRf2y2XRknScfTZ06VKaETCzDVRWX5/eL+pPVqmpmml4RqkRi7m5IPv2NU+3/qRgVMVfN9dFS1KRbXjgfONPLMnG3ZKutbJjh4sHZC4co5jU17uO2MbG5ul+f5L9PlumqsQ+6vjW2S6blk9ymVgycmRqv6jfKWY1J6OYpAuKZsHPglEVYr9hg1vx5qSToi03bHjgQpDLxBLfL5os9vG4iwc0enQ0NhpGNnTv7lZXS/X7tP6kzFSF2Ptj7KP2N48ald2yafmkocGtSpXtxJL6+mP9ovG4iwfUoUM0NhpGtviT/44kzNax/qRgVMXXs2FDfpYoy3bZtHzhTyzJxXcZi7lyFi50nw8ccHGAzB9qlAL19a4PaeVK93n7dutPCkpViH3UY+wTicWcMAYJD5xvophY4vtF/QfY4sXu2swfapQCyYMIFixonm6kp+LF/vBhN5s0X4sPBw0PXAiimFjSvbsbVuqXZZNVjFLizDNd4L7E36f1JwWj4sX+3Xed4OezZg+l4cqJamKJP6T0yBH3fsYZLh6QYRQbkaOLmYB7HzHC+pOCUPFin48x9omEXTYtX0Q5sSQWc+P1V650DxCr1RulRCwGb73l/tuvvGK/z6BUjdjny40DR2vCxQyKFmWgMr+Mhx5y8X/sz2SUEv7v8Re/sP6kMFS82OcjVEIy9fXBlk3LJ9kEP0vHmWe6Fsuvfx1dmYYRFSNGQPv2R3+fJvbBqHixX7/eTbjItOh2LpSC3z7KiSX+YhG7d0PXrm4ii2GUCm3bug7Z3butPykMVSH2+XThQPBl0/JJ1BNL/AfYeefZZBWj9PB/n1arD07rYhuQb6JetCQVfnjghx+GP/0pv+dKx86dMHlydOXZn8koZez3GZ6KF/v16wuzAPGdd7qlCovVSdumDXz5y9GVN3o03HMPTJ0aXZmGERWf/jTcdRdcc02xLSkfKlrs9+51Haf5duOA6zQaMSL/5ykUNTXw7W8X2wrDSE2bNnDzzcW2oryoaG9sIUbiGIZhlAMVLfb5nlBlGIZRLlSF2BfCjWMYhlHKVLTYb9jgRsrYIsSGYVQ7gcReRMaLyGoRWSMit6TY30dEnhWRZSLyvIjUJuy7VkTe9F7XRml8JtavdxMu2rYt5FkNwzBKj4xiLyKtgF8ClwKDgCkiMigp2yzgYVUdAswAfuwd2w24AxgDjAbuEJGu0ZnfMoUYY28YhlEOBKnZjwbWqOpaVT0AzAUmJOUZBPzD234uYf+/AM+o6nZV3QE8A4zP3exg5HPREsMwjHIiiNj3AjYkfN7opSWyFJjkbU8EOolI94DHIiLTRaRRRBq3bt0a1PYWUS1MqATDMIxyIKoO2puBC0XkVeBCYBNwOOjBqjpbVetUta5nz56RGLRtmwt/ajV7wzCMYDNoNwGJ9eNaL60JVd2MV7MXkY7AZ1R1p4hsAsYlHft8DvYGxsbYG4ZhHCVIzX4h0F9ETheRtsBkYF5iBhHpISJ+WT8A7ve2nwY+LSJdvY7ZT3tpecfG2BuGYRwlo9ir6iHgGziRXgk8pqrLRWSGiFzpZRsHrBaRN4CTgJnesduBH+EeGAuBGV5a3rFQCYZhGEcJFAhNVecD85PSbk/Yfhx4PM2x93O0pl8w1q+Hdu3cSvSGYRjVTsXOoPVH4ogU2xLDMIziU7FibxOqDMMwjlKxYm8TqgzDMI5SkWJ/8CBs3mwjcQzDMHwqUuw3b3YzaK1mbxiG4ahIsbcx9oZhGM2pSLG3MfaGYRjNqUixt5q9YRhGcypW7Lt2hY4di22JYRhGaVCRYm9j7A3DMJpTkWJvY+wNwzCaU7Fib/56wzCMo1Sc2O/eDTt3Ws3eMAwjkYoText2aRiGcSwVJ/Y27NIwDONYKk7srWZvGIZxLBUn9uvXQ00NnHpqsS0xDMMoHSpS7E89FVoHWoPLMAyjOqg4sbcJVYZhGMdScWJvY+wNwzCOJZDYi8h4EVktImtE5JYU+3uLyHMi8qqILBORy7z0NiLykIi8JiIrReQHUV9AIkeOWM3eMAwjFRnFXkRaAb8ELgUGAVNEZFBSttuAx1R1ODAZuNdLvxpop6qDgZHA10SkbzSmH8vWrXDggIm9YRhGMkFq9qOBNaq6VlUPAHOBCUl5FOjsbZ8AbE5IP15EWgPHAQeAD3O2Og02xt4wDCM1QcS+F7Ah4fNGLy2RO4HPi8hGYD7wTS/9ceAj4F1gPTBLVbcnn0BEpotIo4g0bt26NdwVJGBj7A3DMFITVQftFOBBVa0FLgMeEZEaXKvgMHAqcDrwXRHpl3ywqs5W1TpVrevZs2fWRvg1exN7wzCM5gQR+01AomOk1ktL5MvAYwCqugBoD/QArgH+qqoHVfV9IA7U5Wp0Otavh+OOg27d8nUGwzCM8iSI2C8E+ovI6SLSFtcBOy8pz3rgYgARORsn9lu99E966ccD5wKrojH9WPyROCL5OoNhGEZ5klHsVfUQ8A3gaWAlbtTNchGZISJXetm+C3xVRJYCvwOmqariRvF0FJHluIfGA6q6LB8XArZoiWEYRjoCBRVQ1fm4jtfEtNsTtlcAsRTH7cENvywI69fDZZcV6myGYRjlQ8XMoP34Y3jvPavZG4ZhpKJixH77djj9dOh3zFgfwzAMo2JiQ55yCqxdW2wrDMMwSpOKqdkbhmEY6TGxNwzDqAJM7A3DMKoAE3vDMIwqwMTeMAyjCjCxNwzDqAJM7A3DMKoAE3vDMIwqwMTeMAyjCjCxNwzDqAJM7A3DMKoAE3vDMIwqwMTeMAyjCjCxNwzDqAJM7A3DMKoAE3vDMIwqwMTeMAyjCggk9iIyXkRWi8gaEbklxf7eIvKciLwqIstE5LKEfUNEZIGILBeR10SkfZQXYBiGYWQm47KEItIK+CXwKWAjsFBE5qnqioRstwGPqeqvRGQQMB/oKyKtgUeBL6jqUhHpDhyM/CoMwzCMFglSsx8NrFHVtap6AJgLTEjKo0Bnb/sEYLO3/WlgmaouBVDVbap6OHezDcMwjDAEEftewIaEzxu9tETuBD4vIhtxtfpveukDABWRp0VksYh8P9UJRGS6iDSKSOPWrVtDXYBhGIaRmag6aKcAD6pqLXAZ8IiI1ODcRPXAVO99oohcnHywqs5W1TpVrevZs2dWBsyZA337Qk2Ne58zJ8srMQzDqECCiP0m4LSEz7VeWiJfBh4DUNUFQHugB64V8IKqfqCqe3G1/hG5Gp3MnDkwfTq88w6ouvfp003wDcMwfIKI/UKgv4icLiJtgcnAvKQ864GLAUTkbJzYbwWeBgaLSAevs/ZCYAURc+utsHdv87S9e126YRiGEWA0jqoeEpFv4IS7FXC/qi4XkRlAo6rOA74L3CciN+E6a6epqgI7RORu3ANDgfmq+peoL2L9+nDphmEY1YY4TS4d6urqtLGxMdQxffs6100yffrAunWRmGUYhlHSiMgiVa1Lt78iZtDOnAkdOjRP69DBpRuGYRgVIvZTp8Ls2a4mL+LeZ8926YZhGEYAn325MHWqibthGEY6KqJmbxiGYbSMib1hGEYVYGJvGIZRBZjYG4ZhVAEm9oZhGFVAyU2qEpGtQIopUoHpAXwQkTmlQKVdD1TeNVXa9UDlXVOlXQ8ce019VDVtJMmSE/tcEZHGlmaRlRuVdj1QeddUadcDlXdNlXY9EP6azI1jGIZRBZjYG4ZhVAGVKPazi21AxFTa9UDlXVOlXQ9U3jVV2vVAyGuqOJ+9YRiGcSyVWLM3DMMwkjCxNwzDqAIqRuxFZLyIrBaRNSJyS7HtiQIRWScir4nIEhEJt6JLCSAi94vI+yLyekJaNxF5RkTe9N67FtPGsKS5pjtFZJN3n5aIyGXFtDEMInKaiDwnIitEZLmIfNtLL8v71ML1lPM9ai8ir4jIUu+a/qeXfrqIvOxp3u+9ZWPTl1MJPnsRaQW8AXwKt8j5QmCKqka+3m0hEZF1QJ2qluVkEBG5ANgDPKyqn/DS/gPYrqr/7j2Uu6rqvxbTzjCkuaY7gT2qOquYtmWDiJwCnKKqi0WkE7AI+G/ANMrwPrVwPZ+lfO+RAMer6h4RaQM0AN8G/jvwR1WdKyK/Bpaq6q/SlVMpNfvRwBpVXauqB4C5wIQi21T1qOoLwPak5AnAQ972Q7g/YtmQ5prKFlV9V1UXe9u7gZVAL8r0PrVwPWWLOvZ4H9t4LwU+CTzupWe8R5Ui9r2ADQmfN1LmN9hDgb+JyCIRmV5sYyLiJFV919t+DzipmMZEyDdEZJnn5ikLl0cyItIXGA68TAXcp6TrgTK+RyLSSkSWAO8DzwBvATtV9ZCXJaPmVYrYVyr1qjoCuBT4uudCqBjU+RDL348IvwLOAIYB7wI/La454RGRjsD/Br6jqh8m7ivH+5Tiesr6HqnqYVUdBtTiPBkDw5ZRKWK/CTgt4XOtl1bWqOom7/194E+4m1zubPH8qr5/9f0i25MzqrrF+zMeAe6jzO6T5wf+38AcVf2jl1y29ynV9ZT7PfJR1Z3Ac8BYoIuI+EvLZtS8ShH7hUB/r3e6LTAZmFdkm3JCRI73OpgQkeOBTwOvt3xUWTAPuNbbvhb4cxFtiQRfFD0mUkb3yev8+w2wUlXvTthVlvcp3fWU+T3qKSJdvO3jcANRVuJE/yovW8Z7VBGjcQC8oVT3AK2A+1V1ZpFNygkR6YerzYNbGP635XZNIvI7YBwuFOsW4A7gCeAxoDculPVnVbVsOjzTXNM4nHtAgXXA1xL83SWNiNQD/wReA454yf8D5+cuu/vUwvVMoXzv0RBcB2wrXAX9MVWd4WnEXKAb8CrweVX9OG05lSL2hmEYRnoqxY1jGIZhtICJvWEYRhVgYm8YhlEFmNgbhmFUASb2hmEYVYCJvWEYRhVgYm8YhlEF/D+wlWoZ/7HiEgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M8yCybRtckRB",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "72cd4cbc-7e0e-4560-f02e-acf927d2b1ed"
      },
      "source": [
        "test_generator = test_datagen.flow_from_directory(\n",
        "        test_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=20,\n",
        "        class_mode='binary')\n",
        "\n",
        "test_loss, test_acc = VGG16_model.evaluate_generator(test_generator, steps=50)\n",
        "print('test acc:', test_acc)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 32 images belonging to 2 classes.\n",
            "test acc: 0.96875\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "t8hZVC20fNGJ",
        "colab_type": "text"
      },
      "source": [
        "---"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "shatxC7bfM1U",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 232
        },
        "outputId": "03596290-1e88-4e26-b107-664dedd801aa"
      },
      "source": [
        "conv_base.trainable = True\n",
        "\n",
        "set_trainable = False\n",
        "for layer in conv_base.layers:\n",
        "    if layer.name == 'block5_conv1':\n",
        "        set_trainable = True\n",
        "    if set_trainable:\n",
        "        layer.trainable = True\n",
        "    else:\n",
        "        layer.trainable = False"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-8-08e17d445f76>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mconv_base\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mtrainable\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mTrue\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0mset_trainable\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0;32mFalse\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0;32mfor\u001b[0m \u001b[0mlayer\u001b[0m \u001b[0;32min\u001b[0m \u001b[0mconv_base\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlayers\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0mlayer\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mname\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m'block5_conv1'\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'conv_base' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3msrsCEwglE3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from keras import models\n",
        "from keras import layers\n",
        "\n",
        "VGG16_FT_model = models.Sequential()\n",
        "VGG16_FT_model.add(conv_base)\n",
        "VGG16_FT_model.add(layers.Flatten())\n",
        "VGG16_FT_model.add(layers.Dense(256, activation='relu'))\n",
        "VGG16_FT_model.add(layers.Dense(1, activation='sigmoid'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IZRl9dxHgWEI",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "88334e78-ca9b-4edc-904e-07a1cdbaad65"
      },
      "source": [
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "\n",
        "train_datagen = ImageDataGenerator(\n",
        "      rescale=1./255,\n",
        "      rotation_range=40,\n",
        "      width_shift_range=0.2,\n",
        "      height_shift_range=0.2,\n",
        "      shear_range=0.2,\n",
        "      zoom_range=0.2,\n",
        "      horizontal_flip=True,\n",
        "      fill_mode='nearest')\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale=1./255)\n",
        "train_generator = train_datagen.flow_from_directory(\n",
        "        train_dir,\n",
        "        # All images will be resized to 150x150\n",
        "        target_size=(150, 150),\n",
        "        batch_size=20,\n",
        "        # Since we use binary_crossentropy loss, we need binary labels\n",
        "        class_mode='binary')\n",
        "\n",
        "validation_generator = test_datagen.flow_from_directory(\n",
        "        validation_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=20,\n",
        "        class_mode='binary')\n",
        "\n",
        "VGG16_FT_model.compile(loss='binary_crossentropy',\n",
        "              optimizer=optimizers.RMSprop(lr=1e-5),\n",
        "              metrics=['acc'])\n",
        "\n",
        "history = VGG16_FT_model.fit_generator(\n",
        "      train_generator,\n",
        "      steps_per_epoch=100,\n",
        "      epochs=30,\n",
        "      validation_data=validation_generator,\n",
        "      validation_steps=50)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 148 images belonging to 2 classes.\n",
            "Found 32 images belonging to 2 classes.\n",
            "Epoch 1/30\n",
            "100/100 [==============================] - 22s 215ms/step - loss: 0.0631 - acc: 0.9817 - val_loss: 0.4494 - val_acc: 0.9062\n",
            "Epoch 2/30\n",
            "100/100 [==============================] - 21s 208ms/step - loss: 0.0079 - acc: 0.9978 - val_loss: 0.1541 - val_acc: 0.8750\n",
            "Epoch 3/30\n",
            "100/100 [==============================] - 20s 203ms/step - loss: 0.0041 - acc: 0.9989 - val_loss: 0.7673 - val_acc: 0.9062\n",
            "Epoch 4/30\n",
            "100/100 [==============================] - 21s 208ms/step - loss: 0.0028 - acc: 0.9984 - val_loss: 1.3237 - val_acc: 0.8750\n",
            "Epoch 5/30\n",
            "100/100 [==============================] - 20s 205ms/step - loss: 0.0012 - acc: 0.9995 - val_loss: 0.5445 - val_acc: 0.8750\n",
            "Epoch 6/30\n",
            "100/100 [==============================] - 21s 208ms/step - loss: 0.0034 - acc: 0.9989 - val_loss: 0.0712 - val_acc: 0.9375\n",
            "Epoch 7/30\n",
            "100/100 [==============================] - 20s 204ms/step - loss: 0.0093 - acc: 0.9978 - val_loss: 0.4627 - val_acc: 0.8750\n",
            "Epoch 8/30\n",
            "100/100 [==============================] - 21s 208ms/step - loss: 0.0072 - acc: 0.9989 - val_loss: 1.0850 - val_acc: 0.8750\n",
            "Epoch 9/30\n",
            "100/100 [==============================] - 20s 203ms/step - loss: 3.2352e-04 - acc: 1.0000 - val_loss: 1.0725 - val_acc: 0.8750\n",
            "Epoch 10/30\n",
            "100/100 [==============================] - 21s 208ms/step - loss: 3.3653e-04 - acc: 1.0000 - val_loss: 1.0298 - val_acc: 0.8750\n",
            "Epoch 11/30\n",
            "100/100 [==============================] - 20s 205ms/step - loss: 0.0015 - acc: 0.9995 - val_loss: 1.0310 - val_acc: 0.8750\n",
            "Epoch 12/30\n",
            "100/100 [==============================] - 21s 208ms/step - loss: 0.0029 - acc: 0.9995 - val_loss: 1.2508 - val_acc: 0.8750\n",
            "Epoch 13/30\n",
            "100/100 [==============================] - 20s 204ms/step - loss: 7.6754e-04 - acc: 0.9995 - val_loss: 1.6062e-06 - val_acc: 0.8750\n",
            "Epoch 14/30\n",
            "100/100 [==============================] - 21s 208ms/step - loss: 0.0017 - acc: 0.9995 - val_loss: 0.4954 - val_acc: 0.8750\n",
            "Epoch 15/30\n",
            "100/100 [==============================] - 21s 209ms/step - loss: 0.0022 - acc: 0.9995 - val_loss: 0.9491 - val_acc: 0.8750\n",
            "Epoch 16/30\n",
            "100/100 [==============================] - 21s 209ms/step - loss: 6.2089e-04 - acc: 1.0000 - val_loss: 0.9494 - val_acc: 0.8750\n",
            "Epoch 17/30\n",
            "100/100 [==============================] - 20s 203ms/step - loss: 8.4630e-05 - acc: 1.0000 - val_loss: 1.0568 - val_acc: 0.8750\n",
            "Epoch 18/30\n",
            "100/100 [==============================] - 21s 208ms/step - loss: 6.5978e-04 - acc: 1.0000 - val_loss: 1.0609 - val_acc: 0.9062\n",
            "Epoch 19/30\n",
            "100/100 [==============================] - 20s 204ms/step - loss: 5.3462e-04 - acc: 0.9995 - val_loss: 1.2852 - val_acc: 0.8750\n",
            "Epoch 20/30\n",
            "100/100 [==============================] - 21s 207ms/step - loss: 1.0870e-04 - acc: 1.0000 - val_loss: 0.3043 - val_acc: 0.9062\n",
            "Epoch 21/30\n",
            "100/100 [==============================] - 20s 204ms/step - loss: 1.1841e-04 - acc: 1.0000 - val_loss: 1.5883 - val_acc: 0.8750\n",
            "Epoch 22/30\n",
            "100/100 [==============================] - 21s 208ms/step - loss: 0.0012 - acc: 0.9995 - val_loss: 1.1276 - val_acc: 0.8750\n",
            "Epoch 23/30\n",
            "100/100 [==============================] - 20s 202ms/step - loss: 5.4458e-04 - acc: 1.0000 - val_loss: 1.5816 - val_acc: 0.9375\n",
            "Epoch 24/30\n",
            "100/100 [==============================] - 21s 211ms/step - loss: 1.6014e-04 - acc: 1.0000 - val_loss: 1.0369 - val_acc: 0.9062\n",
            "Epoch 25/30\n",
            "100/100 [==============================] - 20s 203ms/step - loss: 7.1676e-04 - acc: 0.9995 - val_loss: 8.1810e-05 - val_acc: 0.8750\n",
            "Epoch 26/30\n",
            "100/100 [==============================] - 21s 209ms/step - loss: 0.0011 - acc: 0.9995 - val_loss: 2.1869 - val_acc: 0.8750\n",
            "Epoch 27/30\n",
            "100/100 [==============================] - 20s 205ms/step - loss: 0.0010 - acc: 0.9995 - val_loss: 0.7943 - val_acc: 0.8750\n",
            "Epoch 28/30\n",
            "100/100 [==============================] - 21s 209ms/step - loss: 0.0020 - acc: 0.9995 - val_loss: 1.5652 - val_acc: 0.9375\n",
            "Epoch 29/30\n",
            "100/100 [==============================] - 20s 204ms/step - loss: 2.1491e-04 - acc: 1.0000 - val_loss: 0.6637 - val_acc: 0.8750\n",
            "Epoch 30/30\n",
            "100/100 [==============================] - 21s 208ms/step - loss: 6.5560e-04 - acc: 1.0000 - val_loss: 2.0906 - val_acc: 0.8750\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2uNqKEQbfzd9",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "VGG16_FT_model.save('clean_and_messy_VGG16_FT.h5')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c5I0gN_9ujJ9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "outputId": "e32696d0-ccba-4cf1-b7e3-71c330e0abe5"
      },
      "source": [
        "acc = history.history['acc']\n",
        "val_acc = history.history['val_acc']\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "epochs = range(len(acc))\n",
        "\n",
        "plt.plot(epochs, acc, 'bo', label='Training acc')\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation acc')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.legend()\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "plt.plot(epochs, loss, 'bo', label='Training loss')\n",
        "plt.plot(epochs, val_loss, 'b', label='Validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cA_-J3vsf5fc",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 545
        },
        "outputId": "687127e7-5649-4692-8066-dcecd7f6e667"
      },
      "source": [
        "def smooth_curve(points, factor=0.8):\n",
        "  smoothed_points = []\n",
        "  for point in points:\n",
        "    if smoothed_points:\n",
        "      previous = smoothed_points[-1]\n",
        "      smoothed_points.append(previous * factor + point * (1 - factor))\n",
        "    else:\n",
        "      smoothed_points.append(point)\n",
        "  return smoothed_points\n",
        "\n",
        "plt.plot(epochs,\n",
        "         smooth_curve(acc), 'bo', label='Smoothed training acc')\n",
        "plt.plot(epochs,\n",
        "         smooth_curve(val_acc), 'b', label='Smoothed validation acc')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.legend()\n",
        "\n",
        "plt.figure()\n",
        "\n",
        "plt.plot(epochs,\n",
        "         smooth_curve(loss), 'bo', label='Smoothed training loss')\n",
        "plt.plot(epochs,\n",
        "         smooth_curve(val_loss), 'b', label='Smoothed validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXUAAAEICAYAAACgQWTXAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deZzN9f7A8dfblsTFZSplTVow9pS6QlpQjV8LJYoWKu2L0GZJXS235d6Sq0RKtspNpbQgUsqQVKQksmZSZCxZ5v37432GY8xyZuacOcu8n4/HPOac7/mez/dz5vA+3/P+fj7vj6gqzjnnEkOJaHfAOedc+HhQd865BOJB3TnnEogHdeecSyAe1J1zLoF4UHfOuQTiQd1lS0TeE5Ge4d43mkRklYicHYF2VUSOD9weKSIPhLJvAY7TXUQ+KGg/c2m3rYisDXe7LjpKRbsDLnxEJD3objngL2Bf4P71qjo+1LZUtWMk9k10qnpDONoRkdrAz0BpVd0baHs8EPJ76IonD+oJRFXLZ94WkVXAdar6Udb9RKRUZqBwziUWT78UA5lfr0Wkv4hsBMaISGUReUdE0kTkj8Dt6kHPmS0i1wVu9xKRT0XkicC+P4tIxwLuW0dE5ojINhH5SESeE5FXc+h3KH18SETmBdr7QESqBj1+pYisFpHNInJfLn+fU0Vko4iUDNp2kYgsCdxuKSKfi8gWEdkgIs+KSJkc2horIsOC7vcLPGe9iFyTZd/zReQrEflTRNaIyOCgh+cEfm8RkXQRaZX5tw16/ukiskBEtgZ+nx7q3yY3InJy4PlbROQ7EUkJeqyTiCwNtLlORO4ObK8aeH+2iMjvIjJXRDy+RIH/0YuPo4G/A7WAPth7PyZwvyawE3g2l+efCiwHqgKPAaNFRAqw72vAl0AVYDBwZS7HDKWPVwBXA0cCZYDMIFMfeD7Q/jGB41UnG6r6BbAdOCtLu68Fbu8D7gi8nlZAe6BvLv0m0IcOgf6cA9QDsubztwNXAZWA84EbReT/Ao+dGfhdSVXLq+rnWdr+O/Au8O/Aa3sSeFdEqmR5DYf8bfLoc2ngbeCDwPNuAcaLyImBXUZjqbwKQENgZmD7XcBaIAk4CrgX8BokUeBBvfjIAAap6l+qulNVN6vqG6q6Q1W3AQ8DbXJ5/mpVfUFV9wEvA9Ww/7wh7ysiNYFTgAdVdbeqfgpMy+mAIfZxjKr+oKo7gclAk8D2S4F3VHWOqv4FPBD4G+RkAtANQEQqAJ0C21DVhao6X1X3quoq4L/Z9CM7XQP9+1ZVt2MfYsGvb7aqfqOqGaq6JHC8UNoF+xD4UVVfCfRrAvA9cGHQPjn9bXJzGlAeGB54j2YC7xD42wB7gPoi8jdV/UNVFwVtrwbUUtU9qjpXvbBUVHhQLz7SVHVX5h0RKSci/w2kJ/7Evu5XCk5BZLEx84aq7gjcLJ/PfY8Bfg/aBrAmpw6H2MeNQbd3BPXpmOC2A0F1c07Hws7KLxaRw4CLgUWqujrQjxMCqYWNgX48gp215+WgPgCrs7y+U0VkViC9tBW4IcR2M9tenWXbauDYoPs5/W3y7LOqBn8ABrd7CfaBt1pEPhGRVoHtjwMrgA9EZKWIDAjtZbhw86BefGQ9a7oLOBE4VVX/xoGv+zmlVMJhA/B3ESkXtK1GLvsXpo8bgtsOHLNKTjur6lIseHXk4NQLWBrne6BeoB/3FqQPWAop2GvYN5UaqloRGBnUbl5nueuxtFSwmsC6EPqVV7s1suTD97erqgtUtTOWmvkf9g0AVd2mqnep6nFACnCniLQvZF9cAXhQL74qYDnqLYH87KBIHzBw5psKDBaRMoGzvAtzeUph+vg6cIGI/CNwUXMoef97fw24DfvwmJKlH38C6SJyEnBjiH2YDPQSkfqBD5Ws/a+AfXPZJSItsQ+TTGlYuui4HNqeDpwgIleISCkRuQyoj6VKCuML7Kz+HhEpLSJtsfdoYuA96y4iFVV1D/Y3yQAQkQtE5PjAtZOt2HWI3NJdLkI8qBdfTwOHA78B84H3i+i43bGLjZuBYcAkbDx9dgrcR1X9DrgJC9QbgD+wC3m5ycxpz1TV34K2340F3G3AC4E+h9KH9wKvYSaWmpiZZZe+wFAR2QY8SOCsN/DcHdg1hHmBESWnZWl7M3AB9m1mM3APcEGWfuebqu7GgnhH7O8+ArhKVb8P7HIlsCqQhroBez/BLgR/BKQDnwMjVHVWYfriCkb8WoaLJhGZBHyvqhH/puBcceBn6q5IicgpIlJXREoEhvx1xnKzzrkw8BmlrqgdDbyJXbRcC9yoql9Ft0vOJQ5PvzjnXALx9ItzziWQqKVfqlatqrVr147W4Z1zLi4tXLjwN1VNyunxqAX12rVrk5qaGq3DO+dcXBKRrDOJD+LpF+ecSyAe1J1zLoF4UHfOuQQSU+PU9+zZw9q1a9m1a1feOzsXAWXLlqV69eqULl062l1xrkBiKqivXbuWChUqULt2bXJef8G5yFBVNm/ezNq1a6lTp060u+NcgcRU+mXXrl1UqVLFA7qLChGhSpUq/k3RxbWYCuqAB3QXVf7vz8W7mAvqzjmXyIYOhdmzI9e+B/UsHn74YRo0aECjRo1o0qQJX3zxRcSOtWrVKl577cACO2PHjuXmm28ucHuzZ8/mggsuOGT74sWLmT59er7bW79+PZdeemme+3Xq1IktW7bku/2sVq1aRcOGDQvdjnOxat06GDQI5s6N3DHiOqiPHw+1a0OJEvZ7/PjCtff555/zzjvvsGjRIpYsWcJHH31EjRq5rbZWOFmDeqTkFtT37t2b4/OOOeYYXn/99Tzbnz59OpUqVSpw/5wrLt54w3536RK5Y8RtUB8/Hvr0gdWrQdV+9+lTuMC+YcMGqlatymGHHQZA1apVOeaYYwArazBw4ECaNGlCixYtWLRoEeeddx5169Zl5MiRgI2e6NevHw0bNiQ5OZlJkyblun3AgAHMnTuXJk2a8NRTTwF2dtyhQwfq1avHPffcs79vH3zwAa1ataJZs2Z06dKF9PR0AN5//31OOukkmjVrxptvvnnIa9q9ezcPPvggkyZNokmTJkyaNInBgwdz5ZVXcsYZZ3DllVeyatUqWrduTbNmzWjWrBmfffYZcPCZ89ixY7n44ouz7Vvt2rX57bffWLVqFSeffDK9e/emQYMGnHvuuezcuROABQsW7P/2k/m3yM2uXbu4+uqrSU5OpmnTpsyaZYvofPfdd7Rs2ZImTZrQqFEjfvzxR7Zv3875559P48aNadiw4f6/r3OxZsoUaNgQTjopggdR1aj8NG/eXLNaunTpIdtyUquWqoXzg39q1Qq5iUNs27ZNGzdurPXq1dMbb7xRZ8+eHXS8WjpixAhVVb399ts1OTlZ//zzT920aZMeeeSRqqr6+uuv69lnn6179+7VjRs3ao0aNXT9+vU5bp81a5aef/75+48xZswYrVOnjm7ZskV37typNWvW1F9++UXT0tK0devWmp6erqqqw4cP1yFDhujOnTu1evXq+sMPP2hGRoZ26dLloPaC273pppv23x80aJA2a9ZMd+zYoaqq27dv1507d6qq6g8//KCZ783PP/+sDRo0yLVvmX+btLQ0/fnnn7VkyZL61Vdfqapqly5d9JVXXlFV1QYNGuhnn32mqqr9+/ff326w4OM98cQTevXVV6uq6rJly7RGjRq6c+dOvfnmm/XVV19VVdW//vpLd+zYoa+//rped911+9vZsmVL7m90HvLz79C5UK1bpyqiOmRI4doBUjWX2Bq3Z+q//JK/7aEoX748CxcuZNSoUSQlJXHZZZcxduzY/Y+npKQAkJyczKmnnkqFChVISkrisMMOY8uWLXz66ad069aNkiVLctRRR9GmTRsWLFiQ4/bstG/fnooVK1K2bFnq16/P6tWrmT9/PkuXLuWMM86gSZMmvPzyy6xevZrvv/+eOnXqUK9ePUSEHj16hPxaU1JSOPzwwwGb9NW7d2+Sk5Pp0qULS5cuDblvWdWpU4cmTZoA0Lx5c1atWsWWLVvYtm0brVq1AuCKK6445HlZffrpp/tfz0knnUStWrX44YcfaNWqFY888giPPvooq1ev5vDDDyc5OZkPP/yQ/v37M3fuXCpWrBjy38G5ovLGG3bqGcnUC8Rx+qVmzfxtD1XJkiVp27YtQ4YM4dlnn+WNzCQY7E/LlChRYv/tzPu55abzI7jdkiVLsnfvXlSVc845h8WLF7N48WKWLl3K6NGjC3WcI444Yv/tp556iqOOOoqvv/6a1NRUdu/eHXLfCrJPYVxxxRVMmzaNww8/nE6dOjFz5kxOOOEEFi1aRHJyMvfffz9Dhw4N6zGdC4cpU6BBAzj55MgeJ26D+sMPQ7lyB28rV862F9Ty5cv58ccf999fvHgxtWrVCvn5rVu3ZtKkSezbt4+0tDTmzJlDy5Ytc9xeoUIFtm3blme7p512GvPmzWPFihUAbN++nR9++IGTTjqJVatW8dNPPwEwYcKEbJ+f13G2bt1KtWrVKFGiBK+88gr79u0L+TWHolKlSlSoUGH/SKKJEyfm+ZzWrVszPnCB5IcffuCXX37hxBNPZOXKlRx33HHceuutdO7cmSVLlrB+/XrKlStHjx496NevH4sWLQpr/50rrA0b4NNPI3+WDjFWJiA/une33/fdZymXmjUtoGduL4j09HRuueUWtmzZQqlSpTj++OMZNWpUyM+/6KKL+Pzzz2ncuDEiwmOPPcbRRx+d4/YqVapQsmRJGjduTK9evahcuXK27SYlJTF27Fi6devGX3/9BcCwYcM44YQTGDVqFOeffz7lypWjdevW2Qbvdu3aMXz4cJo0acLAgQMPebxv375ccskljBs3jg4dOhx0Fh8uo0ePpnfv3pQoUYI2bdrkmSLp27cvN954I8nJyZQqVYqxY8dy2GGHMXnyZF555RVKly7N0Ucfzb333suCBQvo168fJUqUoHTp0jz//PNh779zhVFUqReI4hqlLVq00KyLZCxbtoyTI/3dxEVFeno65cuXB2D48OFs2LCBZ555Jsq9yp7/O3Th1qYN/PYbfPdd4dsSkYWq2iKnx/NMv4jISyKySUS+zeHx7iKyRES+EZHPRKRxYTrsEtO7775LkyZNaNiwIXPnzuX++++PdpecKxIbNthko6I4S4fQ0i9jgWeBcTk8/jPQRlX/EJGOwCjg1PB0zyWKyy67jMsuuyza3XCuyL35ZtGlXiCEoK6qc0Skdi6PfxZ0dz5QvfDdcs65xDBlio14adCgaI4X7tEv1wLv5fSgiPQRkVQRSU1LSwvzoZ1zLrZs3Ahz5hTdWTqEMaiLSDssqPfPaR9VHaWqLVS1RVJSUrgO7ZxzMamoUy8QpiGNItIIeBHoqKqbw9Gmc87FuylTrM5LUaVeIAxn6iJSE3gTuFJVfyh8l6IrEUvvFqadadOmMXz48Gz3yxyimJMtW7YwYsSI/fdDLeUbirZt25J1SKxz4ZSRYfNgClr7/NdfD6ReinLtlVCGNE4APgdOFJG1InKtiNwgIjcEdnkQqAKMEJHFIhK3/9MStfRuYaSkpDBgwIACPTdrUA+1lK9zseDf/4ZHHrGgXJBLgG++aR8MRZl6gRCCuqp2U9VqqlpaVaur6mhVHamqIwOPX6eqlVW1SeAnx0HxsS4RS++ClRn4LmjWQ+ZZ7pdffkmrVq1o2rQpp59+OsuXLz/kucHfHn7++WdatWq1v8ZKpvT0dNq3b0+zZs1ITk7mrbfe2v/6fvrpp/3ldoNL+eZUWje3Er85mTBhAsnJyTRs2JD+/e2Szr59++jVq9f+v3nm3/ff//439evXp1GjRlx++eV5tu2Kp6VLYcAAOP102LoVbrkl/21MmQInnmildotUbiUcI/mTV+nd225TbdMmvD+33ZZjNUtVTdzSu08++aQ++OCDqqq6fv16PeGEE1RVdevWrbpnzx5VVf3www/14osvVlU9qF/BZXsvvPBCffnll1VV9dlnn9UjjjhCVVX37NmjW7duVVXVtLQ0rVu3rmZkZBxUSlc1tNK6uZX4DdamTRtdsGCBrlu3TmvUqKGbNm3SPXv2aLt27XTq1KmampqqZ5999v79//jjD1VVrVatmu7ateugbVl56d3i7a+/VJs1U01KUt24UXXYMCvr/frrobfx66+qJUqo3n9/+PtHopbejYRELb3btWvX/WmPyZMn789rb926lS5dutCwYUPuuOOOg87mszNv3jy6desGwJVXXrl/u6py77330qhRI84++2zWrVvHr7/+mmtbOZXWzelvkJMFCxbQtm1bkpKSKFWqFN27d2fOnDkcd9xxrFy5kltuuYX333+fv/3tbwA0atSI7t278+qrr1KqVNyWPnLZ+OILWL++8O089BAsWgSjRsFRR8E990CzZtC3r031D0W0Ui8QwwW9nn46OsfNLL3btm1bkpOTefnll+nVqxcQ/dK7WaswLl68OKQ2jz32WKpUqcKSJUuYNGnS/nTRAw88QLt27Zg6dSqrVq2ibdu2ebYl2VzxGT9+PGlpaSxcuJDSpUtTu3Ztdu3aFVLfshOO8r2VK1fm66+/ZsaMGYwcOZLJkyfz0ksv8e677zJnzhzefvttHn74Yb755hsP7gkgPR3OPBOOPNKm5NeuXbB25s+3PHqvXvB//2fbSpeGMWOgRQu49VYI5TLYlClwwgmQnFywfhSGn6kHSdTSu2DT9B977DG2bt1Ko0aNADtTP/bYYwEO+kaSkzPOOGN/2dzxQesGbt26lSOPPJLSpUsza9as/WfWub2+nErr5lfLli355JNP+O2339i3bx8TJkygTZs2/Pbbb2RkZHDJJZcwbNgwFi1aREZGBmvWrKFdu3Y8+uijbN26df+1CRff5s6F3bttss9ZZ9kCz/m1fTtceSXUqAFZa801agQPPAATJsDUqbm3s2mTjZgp6lEvmTyoB0lPT6dnz577L6QtXbqUwYMHh/z8iy66iEaNGtG4cWPOOuusg0rvZre9UaNG+0vvZl7Iy05w6d1GjRrRqlUrvv/+e8qWLbu/9G6zZs048sgjc2zj0ksvZeLEiXTt2nX/tnvuuYeBAwfStGnTkM6Gn3nmGZ577jmSk5NZF/S/pnv37qSmppKcnMy4ceM4KbAAY5UqVTjjjDNo2LAh/fr1O6itvn37kpGRQXJy8v40V/AZeqiqVavG8OHDadeuHY0bN6Z58+Z07tyZdevW0bZtW5o0aUKPHj345z//yb59++jRo8f+i7O33nqrL5gdRqoQwjlKRMycCWXKwEcfWYqkfXsbUpgf/frBTz/Byy9DIFt3kAEDoEkTuPFG2JzLbJypU6OXegFi90Kpc9Hi/w4L5tFH7YJi/fqqt9yiOnWq6u+/F82xmzZVbdvWbs+dq1qunGrDhqq//Rba8997z/p+112577d4sWqpUqo9euS8T/v2qvXqqWZkhHbs/MIvlDrnIu2XX2DwYDj1VEtfjB4NF10EVapYLrp/f5gxw1Ic4bZ5MyxebGkXgH/8A6ZNgx9/hHPPhS1b8n7+NdfYrM9hw3Lft3Fjm5D06qt2jKzS0mDWrOilXsDTL865MLj7bku/TJoE778Pf/xhsykHDbJlJp96Cjp0gMqVoXVr+wDILYWRH7Nn27Hbtz+wrX17G4HyzTfQqVPOaSHVA6NaXnkFypbN+3j33ms59uuvh99/P/ixqKdeiMGgrlFaick58H9/BTFzpo32GDgQMscVlCljwXvQIAvuf/xhZ+p33gl//QVDh8Idd4Tv+OXLwymnHLy9UyeYOBG+/BJSUmDHjkOfO2ECTJ5sHzJNm4Z2vDJlYOxYOyvP+hqmTIHjj7cz+qjJLTcTyZ/scuorV67UtLQ0zYhUMsq5XGRkZGhaWpquXLky2l2JG7t3Ww69Th3VHTtCf95VV6lWrqwamPtWKCeeqNqpU86Pjx+vKqJ63nmqgXlnqqq6Zo1qpUqqrVoVrB8PPGB5+LfftvtpaaolS6oOHJj/tvKDPHLqMTVAt3r16qxduxavte6ipWzZslSv7uu8hOq552xK/f/+B4cfHvrzOneGcePg008hhOkROVq3DpYvh969c97niitg1y649lq47DI7my5ZEq6+GvbssX4UZKrC/ffb677+elt7dOpU2LcvuqkXiLHJR6VLl6ZOnTrR7oZzLgS//mrplQ4dLL2RH+eea2mMadMKF9RnzrTfwfn07FxzDezcCTffDD16QKtWNvxx5EhLlxREmTI2KenUUy2ttHYt1K1rwx6jKeZy6s65+DBggAXKZ57J/0iP8uUtEE+bZhcrC+rjj22ETWA+Xa5uugkef9xy6HfcAR07Qp8+BT82QPPm9ncYM8Y+JKI56iWTB3XnXL59/rldLLzzTpsOXxApKTbZZ9mygj1f1c7U27WDEiFGsrvvhn/+09YMHT06PAH4gQdsOKQW8QpHOfGg7pzLl337LI1xzDGWVy6oCy+039mN9w7FihWwZs2B8emhGjDArgNUq1aw42Z12GGWpx8+PPQRNJHkQd05ly+jR1sVwyeesDRKQR17rE1MCpTfz7dQ8+lF4eSTbYJVtFMv4EHdOZcPv/9uk2/OPBPCscZISoqVzN24Mf/P/fhjqF4d6tUrfD8SiQd151zIHnjAJhL95z/hOStNSbFc9Lvv5u95GRk2Hf+ss2Lj7DiWeFB3zoXkq69sCOBNN4U22iQUjRrZLNT85tW/+eZANUZ3MA/qzrk8qdo6nVWq2BT/cBGxs/UPP8x+Gn9OMvPp+b1IWhx4UHfO5Wn8eJg3z4YDhrsEfUqKjXf/6KPQn/PxxzaU0if/HsqDunMuV3/+aQtInHKKTa0PtzPPtEUpQk3B7NkDn3ziZ+k5yTOoi8hLIrJJRL7N4XERkX+LyAoRWSIizcLfTedcNOzbZ6NdNm6EZ58NfZJPfpQpYxUV337bLoDmJTXV1iT1fHr2QnmLxgIdcnm8I1Av8NMHeL7w3XLORdMff9g49Lp1rWjX9ddDy5aRO15Kiq3t+cUXee/78cf2uzA1YxJZnkFdVecAv+eyS2dgXKAq5HygkoiEaa6Wc64off+9LRpRvbqlXGrXtsUmnnsussft2NEqJYaSgpk504pmVa0a2T7Fq3B8mToWWBN0f21g2yFEpI+IpIpIqpfXdS42ZGTAe+9ZtcWTT4aXXrIStV99ZasKXXSRlaqNpEqVoE2bvIP6zp3w2WeeT89NkV4oVdVRqtpCVVskJSUV5aGdc1mkp9sZeP36ltNesgQeesjWG33ppaIvIZuSYjVZVqzIeZ/PPrOVkzyfnrNwBPV1QI2g+9UD25xzMWrOHEux3HwzVKxoQxZXrbICXUceGZ0+hVLg6+OPLU3TunXR9CkehSOoTwOuCoyCOQ3YqqobwtCucy5C7r/fhhF+/rldnLziChuFEk116kByct5BvWVLqFCh6PoVb0IZ0jgB+Bw4UUTWisi1InKDiNwQ2GU6sBJYAbwA9I1Yb51zhbZgAcyda7XQTzst2r05WOfOtsTd5s2HPrZ1qw1n9NRL7kIZ/dJNVaupamlVra6qo1V1pKqODDyuqnqTqtZV1WRVTY18t50rPr7/Hh55BPbuDU97Tz1lZ+nXXBOe9sIpJcXGxr/33qGPffKJXdT1i6S58xmlzsWwTZvgvPPgvvss711Ya9bYcm69e1tgjzXNm9viFdmlYGbOhLJlbX1RlzMP6s7FqL/+gosvtsBerx4MGQK7dxeuzf/8x37fckvh+xcJJUrYBdP33rPXH+zjj+Ef/7CVhlzOPKg7F4NUbVHkefPg5Zdtceeff7YFjgtq2zYYNQouvdTK3caqzp1tuOXs2Qe2/forfPut59ND4UHduRj0+OMwbhwMHgxdu9rEoNNPt3Hku3YVrM2XXrKLjXfeGdauht1ZZ0G5cgenYGbNOvCYy50HdedizLRptjhy167w4IO2TQSGDYN16+C//81/m/v2wdNPW/oikjVcwqFsWbuOMG2afWMBy6dXrAjNvFxgnjyoOxdDvvkGune3C4Zjxhy8VFu7dnam+sgjsH17/tr93/9sclGsn6VnSkmBtWutVAFYPr1NG5t45HLnQd25GLFpk10krFDBgnC5cofu89BDtl/mBc9QPfkkHHecBct4cP75dtF02jT7MFq50vPpofKg7qImIwMGDrQ62sVd5kiXX3+1QHZstiXxLK/eqRM89pjlx0Mxf77VTLn99sgX5gqXpCR7rdOmHVi6zoN6aDyou6h56CEYPhzuuutA7rQ4UrV65ZkjXVq0yH3/hx6yeudPPRVa+089ZfnoSKxaFEkpKZZ+GTsWjjrKCo+5vHlQd1Hxzjs2suP44+HHH60GSXH1xBMWzAcNsoujeWnWDC65xFIq2U2nD7ZqFbz+un1olC8flu4WmcxU0dy5di0h+PqCy5kHdVfkfvwRevSw4PTZZ5Y7Hjs22r2Kjrffhv79Dx7pEoohQ2ws9+OP577ff/5juelYnWyUmxNPtB/woYz54UHdFan0dPi//7NRDG++abnTSy+FSZNsAYTi5JtvrDpi5kiX/Kz/2aCBPfff/7b1Q7Pz55/wwgv2gVG9enj6XNQyz9Y9nx46D+quyKhaXvf772HixAOzGnv2tAD01lvR7V9Reu89OPfc3Ee65GXQICsbMHx49o+PHm2zSO+4o3B9jaZ777VUXZ060e5J/PCg7orM449bfnf4cDj77APb27aFmjUtr5zotm6F666zESxVqsAHH+Q80iUv9epBr17w/PNWqCvY3r1WWuDMM/O+8BrLKlWy4Y0udB7UXZH48EMbvti1K9x998GPlSgBV15pAW79+uj0ryh8+KEtAjFmjP0tFi6Ehg0L1+YDD9g3oIcfPnj71KmwenX8TDZy4eNB3UXcqlVw+eU2JG306OxHMfTsaePWX321yLsXcdu2wQ03WLrliCNspM8jj4Sn2mCtWlb4a/Rom6ADFuT/9S8bWXTBBYU/hosvHtRdRO3caZNq9u2zs8echtXVq2eTTV5+ObHGrM+caWfno0ZBv3427jrctVfuvdcuPA8davczl6iLp8lGLnw8qLuIyZxUs3ixLfBw/PG57wMR2CgAABd5SURBVN+zp60mv3Bh0fQvktLTbVHn9u1t7c9PP7VZoGXLhv9YxxwDN90Er7xiF6GffBIqV7Z8uyt+PKi7iHn2WQs0gweHdrGra1dLScT7BdM5c6BxYxgxws6WFy+2byGR1L8/HH64fYhOnWq/jzgissd0scmDuouIzIWNU1Js5fpQVKpkY9hfe+3QVW/iQUaGfYC1aWP3Z8+2KfoFGa6YX0lJ9gEyZ45deL755sgf08UmD+ou7H7+Gbp0saqA48blb1JNr17w++/w7rsR615EbN9u3zSGDLE00pIlNpywKN11F/z971a6t6DDJF3886Duwuqdd2yG5K5dlgaoWDF/zz/nHFt4OJ5SML/8YotPTJ1q+ewxY6KT+qhc2a5JPP980R/bxY6QgrqIdBCR5SKyQkQGZPN4TRGZJSJficgSEekU/q66WLZ3r43CuPBCG2a3cGHBquqVLGl1YaZPh7S08Pcz3ObNg1NOseGE775rszejWXjqqKMst+6KrzyDuoiUBJ4DOgL1gW4ikvW/6/3AZFVtClwOjAh3R13s2rjRZoj+85/Qu7cV6apbt+Dt9expHxKvvRa+PkbCmDG2GtHf/mZDCDt0iHaPnAvtTL0lsEJVV6rqbmAi0DnLPgr8LXC7IpDA8wJdsE8+gaZN4csvLWUyalThzxQbNLCp7bGagtm71y4CX3ONXRT94gs46aRo98o5E0pQPxYIriyxNrAt2GCgh4isBaYD2Rb6FJE+IpIqIqlp8fDd2uUoI8NquJx1luXNv/wSrroqfO337GkTdZYsCV+b4bBli83SfOopK2f73nt2cdK5WBGuC6XdgLGqWh3oBLwiIoe0raqjVLWFqrZISkoK06FdUfv9d+jc2eqXdOkCCxYUvoZJVt26QenSsXW2/uOPcNpptgjyqFFW9tYXQnaxJpSgvg6oEXS/emBbsGuByQCq+jlQFqgajg662JKaaotbzJhhk4smTLDyseFWpYqdEY8fb+mOaPvgA5vev3mzBfXevaPdI+eyF0pQXwDUE5E6IlIGuxA6Lcs+vwDtAUTkZCyoe34lwYwaBWecYdP/P/3UpqZHcqRHr162EPOMGZE7Rl62bLGCWeedBzVqWJqpqMefO5cfeQZ1Vd0L3AzMAJZho1y+E5GhIhJYl4S7gN4i8jUwAeilmkhlmYq3zNKu119vOfRFi8JflCo7HTvaTMlopGBUYcoUOPlkq4B4991WKMsXa3CxLqSMoKpOxy6ABm97MOj2UuCM8HbNxQJVy50/+qjVPH/ppaLLI5cubUu2Pf+85fGL6oLkmjXQt69NpGrWzMafN2tWNMd2rrB8RqnLUUaGjfB49FGrBz52bNFfGOzZ05ZsmzQp8sfat88uftavbyVz//UvG67oAd3FEw/qLlt798K118Jzz1nqYcSI/NVwCZcmTaBRo8inYJYssUqKt91m1w2+/dbGovvoFhdvPKi7Q+zebWmPsWOtQNVjj0Vv6ruIna1/8QUsXx7+9nfutPRS8+ZWiGz8eBt77rlzF688qLuD7NwJF11kFwmfeAIefDC6tUzAqg6WLBn+s/UvvrBViYYPt3ozy5bZh1m0X69zheFB3e23bZstZvHee/Df/1op11hw1FFWV2XcOPjzz/C0OWoUtG5tefSPP7Y6LlWqhKdt56LJg7oD4I8/bGHkOXNstaI+faLdo4Pdeits2GArCn3yScHb2bXLJg5lDs9cuNB+O5coPKg70tIOjD+fMsXSHbHm3HNtNaWSJa0y4l13WYDOjzVrrADXiy/CfffZUEWv2+ISjQf1Ym7jRpshuXw5TJtm+fRYdfrp8PXXcOONthhF8+ahL1I9e7btv2wZvPkmDBtmHxDOJRoP6sXYvn1WOOuXX+D9920qfKw74ggbZvn++zaF/7TTYOhQ2LMn+/1VraLi2WfbWfmXX8b2B5dzheVBvRh76CE7gx0xIv7qmZx3no0lv+wyGDTIxpZ///3B++zYYaNa7rzTVmT68kuve+4Snwf1YmrmTDvDveoqGwcejypXhldfhcmTbTm5pk3hmWdsJuzKldCqlVWRHDYM3njDVihyLtFJtOputWjRQlNTU6Ny7OJu0yYbRVKxopXSLV8+2j0qvI0bbVTLO+/YWfvSpbb9tdd8mTmXWERkoaq2yOlxP1MvZjIyrDDXH3/YGW4iBHSAo4+2C70vvmgXU2vUsA8sD+iuuPHKFsXMY4/Zgg8jR1pNlUQiYvVqLr0UypWzKo/OFTce1IuRefPg/vuha9fYm1wUThUrRrsHzkWPp1+Kic2bbfhi7drwwgte38S5ROVn6sWAKlx9tV1M/PxzHwXiXCLzoF4MPP00vP22Dfdr3jzavXHORZIH9XyaOtUKS517LtStG/tpjC+/hP79oXNnW8XIOZfYPKjnw5YtVuxq5067X6eOzWw891wriBVrF+i2bIHLL4dq1Wxt0Vj/AHLOFZ4H9XwYP94C+htvWH56xgyb0ThypBWHOvXUA0G+RYvoLoWmapNx1qyxcrpejdC54sFnlIZI1aahlyhhJWoz7dkD8+fb2O8ZM2zCiypUqmT1Rp5/3opQFbXnn4e+fW1cer9+RX9851xkhGVGqYh0EJHlIrJCRAbksE9XEVkqIt+JyGsF7XCsSk21mYq9ex+8vXRpW0HnoYcsf52WZivfX3SRLTYxZEjR93XtWlssukOH2Fm9yDlXNPJMEIhISeA54BxgLbBARKap6tKgfeoBA4EzVPUPETkyUh2OlhdesFmKV1yR+35Vqtjknq5dLSXz5JM2LT85uWj6CTBggJXVHTHCvlk454qPUP7LtwRWqOpKVd0NTAQ6Z9mnN/Ccqv4BoKqbwtvN6EpPt2p/Xbvm72Lo8OFWSfCGG6zmSlH47DPL/d99t13Idc4VL6EE9WOBNUH31wa2BTsBOEFE5onIfBHJtoySiPQRkVQRSU1LSytYj6Ng4kQL7FlTL3mpUgWeeMIC7UsvRaZvwTIy4Lbb4Jhj7GzdOVf8hOvLeSmgHtAW6Aa8ICKVsu6kqqNUtYWqtkhKSgrToSPvhRegfn2rz51fV11l62Lec4+VvI2kceMs9//oo4lTfdE5lz+hBPV1QI2g+9UD24KtBaap6h5V/Rn4AQvycW/JErsA2rt3wcZ5i9hIlPT0yI5C+fNPOztv1So2F452zhWNUIL6AqCeiNQRkTLA5cC0LPv8DztLR0SqYumYlWHsZ9S8+CKUKWMXOwvq5JPtTH3cOJg1K3x9C/bII/Drr1YKwCcZOVd85RnUVXUvcDMwA1gGTFbV70RkqIikBHabAWwWkaXALKCfqm6OVKeLys6dNizxkkssP14Y990Hxx0HN94If/0Vnv5lWrHCFlfu1QtOOSW8bTvn4ktIOXVVna6qJ6hqXVV9OLDtQVWdFritqnqnqtZX1WRVnRjJTheVN96wqfb5vUCancMPh+eeg+XL7eJpON19t32beOSR8LbrnIs/Poo5Fy+8AMcfD23bhqe9Dh2gSxdbCPmnn8LT5ocfwltv2eIX1aqFp03nXPzyoJ6D5cutZsp114U3R/300zYL9eabrZxAYezdC7ffbtUib789PP1zzsU3D+o5ePFFK8jVq1d42z3mGHj4YXj/fZgypXBtjRwJS5fCv/4Fhx0Wnv455+KbF/TKxu7dUL261XR5443wt79vn1V0XL8eli0rWMnezZuhXj1b9OKDD3zEi3PFRVgKehU3b71lhbnCcYE0OyVL2ln2xo3wwAMFa2PQIBub/tRTHtCdcwd4UM/GCy9AzZpwzjmRO0aLFnDTTTYiJr9fWL75xiY03XgjNGwYmf455+KTB/Usfv7ZRpRce62dUUfSsGFw5JFW8GvfvtCeo2oXRStVik5ZX+dcbPOVj7IYPdrK1V5zTeSPVbGijYa5/HK49FLLs9ete+Anu1z7W2/BzJnw7LO+mpFz7lB+oTTI3r1Qq5atcPTOO0VzTFVbEHrKlEMLflWteiDAH3+8/R4yxCYyLV4c3eXynHPRkdeFUg8LQaZPtxEpI0YU3TFF7Kz72Wdh2zZYudKm/f/004GfefOs/G9mTfYPP/SA7pzLnoeGIC+8YLMyzz8/OsevUAEaN7afrHbvhlWrrB5Ndo875xx4UN9v7Vo7Ux8wIDbPgsuUgRNOiHYvnHOxzke/BIwZY+mNa6+Ndk+cc67gPKhjwXz0aGjf3srjOudcvPKgji3UvHp15GaQOudcUSn2QX32bKvEePrpcNFF0e6Nc84VTtwF9fR0eOwxG1NeWEuWQOfONv777bftYqRzzsWzuAvqb7wB/fvbWfX27QVvZ/Vq6NgRype3Mrg+O9M5lwjiLqj37GlFsKZPtxWJNm7MfxubN9sqRNu3W0CvWTPs3XTOuaiIu6AO0Lev1UBZuhRatbKa5KHasQMuvNAKd731FiQnR66fzjlX1OIyqANccAF88onNsDz9dLudl717rXjW/Pk24qVNm8j30znnilLcBnWwmuTz58PRR8O558Jrr+W8r6rVH3/7bauzcsklRddP55wrKiEFdRHpICLLRWSFiAzIZb9LRERFJMcKYuFWuzZ89pmlYbp3h+HDs1/QefBgW3f03nstfeOcc4koz6AuIiWB54COQH2gm4jUz2a/CsBtwBfh7mReKleGGTPgiitg4EBbdCJ4yON//wtDh9oi0sOGFXXvnHOu6IRypt4SWKGqK1V1NzAR6JzNfg8BjwK7wti/kB12GLzyip2JjxoFKSlWyvZ//7Mz806dbLuv5+mcS2Sh1CM8FlgTdH8tcGrwDiLSDKihqu+KSL+cGhKRPkAfgJoRGEdYogQ8/LAtdNG3r6VkfvrJcu+TJ0Pp0mE/pHPOxZRCXygVkRLAk8Bdee2rqqNUtYWqtkhKSirsoXPUp49dEF292sagv/suHHFExA7nnHMxI5Qz9XVAjaD71QPbMlUAGgKzxXIbRwPTRCRFVaO2Xl3HjrB8uQXz7Nb6dM65RBRKUF8A1BOROlgwvxy4IvNBVd0KVM28LyKzgbujGdAzHXNMtHvgnHNFK8/0i6ruBW4GZgDLgMmq+p2IDBWRlEh30DnnXOhCWrhNVacD07NsezCHfdsWvlvOOecKIq5nlDrnnDuYB3XnnEsgHtSdcy6BeFB3zrkE4kHdOecSiAd155xLIB7UnXMugXhQd865BOJB3TnnEogHdeecSyAe1J1zLoF4UHfOuQTiQd055xKIB3XnnEsgHtSdcy6BeFB3zrkE4kHdOecSiAd155xLIB7UnXMugXhQd865BOJB3TnnEogHdeecSyAhBXUR6SAiy0VkhYgMyObxO0VkqYgsEZGPRaRW+LvqnHMuL3kGdREpCTwHdATqA91EpH6W3b4CWqhqI+B14LFwd9Q551zeQjlTbwmsUNWVqrobmAh0Dt5BVWep6o7A3flA9fB20znnXChCCerHAmuC7q8NbMvJtcB72T0gIn1EJFVEUtPS0kLvpXPOuZCE9UKpiPQAWgCPZ/e4qo5S1Raq2iIpKSmch3bOOQeUCmGfdUCNoPvVA9sOIiJnA/cBbVT1r/B0zznnXH6Ecqa+AKgnInVEpAxwOTAteAcRaQr8F0hR1U3h76ZzzrlQ5BnUVXUvcDMwA1gGTFbV70RkqIikBHZ7HCgPTBGRxSIyLYfmnHPORVAo6RdUdTowPcu2B4Nunx3mfjnnnCsAn1HqnHMJxIO6c84lEA/qzjmXQDyoO+dcAvGg7pxzCcSDunPOJRAP6s45l0A8qDvnXALxoO6ccwnEg7pzziUQD+rOOZdAPKg751wC8aDunHMJJK6C+vjxULs2lChhv8ePj3aPnHMutoRUejcWjB8PffrAjsDy1qtX232A7t2j1y/nnIslcXOmft99BwJ6ph07bHt2/KzeOVccxU1Q/+WX0LdnntWvXg2qB87qswvsHvydc4kkboJ6zZqhbw/1rD4/wT9zf/8AcM7FsrgJ6g8/DOXKHbytXDnbnlWoZ/X5Sen42b9zLh7ETVDv3h1GjYJatUDEfo8alf1F0lDP6vOT0on22b9/UDjnQqKqUflp3ry5Rsqrr6qWK6dqYdV+ypWz7cFq1Tp4n8yfWrUObVMk+31FCt5mqP0Mdb/g/WvVsr7VqlX4/ZxzsQNI1Vxia0IGddXQAlZ+gmWowTrU4J+fNhP1gyLR2nSuKIQlqAMdgOXACmBANo8fBkwKPP4FUDuvNiMd1EOVnyAQrbP/RPygSLQ2M/eNhw8fbzP228xNoYM6UBL4CTgOKAN8DdTPsk9fYGTg9uXApLzajZWgnh/ROvtPxA+KRGszXj58vM3YbzMv4QjqrYAZQfcHAgOz7DMDaBW4XQr4DZDc2o3HoB6qcJ/9J+IHRaK1GS8fPt5m7LeZl3AE9UuBF4PuXwk8m2Wfb4HqQfd/Aqpm01YfIBVIrVmzZv5eSYIK91e8ePmgSLQ24+XDx9uM/TbzElNBPfgnkc/Uoy0ePigSrc14+fDxNmO/zbx4+sWFRaJdiAp3m/Hy4eNtxn6beQlHUC8FrATqBF0obZBln5uyXCidnFe7HtRdoomHDx9vMz7azE1eQV1sn9yJSCfg6cBImJdU9WERGRpofJqIlAVeAZoCvwOXq+rK3Nps0aKFpqam5nls55xzB4jIQlVtkdPjIdVTV9XpwPQs2x4Mur0L6FLQTjrnnAuPuKn94pxzLm8e1J1zLoF4UHfOuQTiQd055xJISKNfInJgkTRgdQGfXhUbC59IEu01JdrrgcR7TYn2eiDxXlN2r6eWqibl9ISoBfXCEJHU3Ib0xKNEe02J9nog8V5Tor0eSLzXVJDX4+kX55xLIB7UnXMugcRrUB8V7Q5EQKK9pkR7PZB4rynRXg8k3mvK9+uJy5y6c8657MXrmbpzzrlseFB3zrkEEndBXUQ6iMhyEVkhIgOi3Z9wEJFVIvKNiCwWkbgrXSkiL4nIJhH5Nmjb30XkQxH5MfC7cjT7mF85vKbBIrIu8D4tDlQvjQsiUkNEZonIUhH5TkRuC2yPy/cpl9cTz+9RWRH5UkS+DrymIYHtdUTki0DMmyQiZXJtJ55y6iJSEvgBOAdYCywAuqnq0qh2rJBEZBXQQlXjctKEiJwJpAPjVLVhYNtjwO+qOjzw4VtZVftHs5/5kcNrGgykq+oT0exbQYhINaCaqi4SkQrAQuD/gF7E4fuUy+vpSvy+RwIcoarpIlIa+BS4DbgTeFNVJ4rISOBrVX0+p3bi7Uy9JbBCVVeq6m5gItA5yn0q9lR1DlZHP1hn4OXA7Zex/3BxI4fXFLdUdYOqLgrc3gYsA44lTt+nXF5P3AqsgZEeuFs68KPAWcDrge15vkfxFtSPBdYE3V9LnL+RAQp8ICILRaRPtDsTJkep6obA7Y3AUdHsTBjdLCJLAumZuEhVZCUitbEFbb4gAd6nLK8H4vg9EpGSIrIY2AR8iK33vEVV9wZ2yTPmxVtQT1T/UNVmQEfgpsBX/4QRWIIrfvJ8OXseqAs0ATYA/4pud/JPRMoDbwC3q+qfwY/F4/uUzeuJ6/dIVfepahOgOpaZOCm/bcRbUF8H1Ai6Xz2wLa6p6rrA703AVOzNjHe/BvKemfnPTVHuT6Gp6q+B/3QZwAvE2fsUyNO+AYxX1TcDm+P2fcru9cT7e5RJVbcAs4BWQCURyVylLs+YF29BfQFQL3A1uAy2yPW0KPepUETkiMCFHkTkCOBc4NvcnxUXpgE9A7d7Am9FsS9hkRn8Ai4ijt6nwEW40cAyVX0y6KG4fJ9yej1x/h4liUilwO3DsQEhy7Dgfmlgtzzfo7ga/QLZL4Id5S4Viogch52dg60Z+1q8vSYRmQC0xcqE/goMAv4HTAZqYiWWu6pq3Fx4zOE1tcW+1iuwCrg+KB8d00TkH8Bc4BsgI7D5XiwPHXfvUy6vpxvx+x41wi6ElsROuCer6tBAjJgI/B34Cuihqn/l2E68BXXnnHM5i7f0i3POuVx4UHfOuQTiQd055xKIB3XnnEsgHtSdcy6BeFB3zrkE4kHdOecSyP8D23nAxuprt7cAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4e056I4_gDxZ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "fa99e58c-0770-40c4-9b49-f25d0e5c7fe2"
      },
      "source": [
        "test_generator = test_datagen.flow_from_directory(\n",
        "        test_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=20,\n",
        "        class_mode='binary')\n",
        "\n",
        "test_loss, test_acc = VGG16_FT_model.evaluate_generator(test_generator, steps=50)\n",
        "print('test acc:', test_acc)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 32 images belonging to 2 classes.\n",
            "test acc: 0.96875\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2tS4-O4GjrWM",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "07Qc7gkcYPki",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import numpy as np"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mPmSnDl32BYh",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 68
        },
        "outputId": "d6f90794-a875-4f4a-ba6e-438d3c576e28"
      },
      "source": [
        "ytest_dir = '/content/messy_vs_clean_room/ytest'\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "ytest_datagen = ImageDataGenerator(rescale=1./255)\n",
        "ytest_generator = ytest_datagen.flow_from_directory(\n",
        "        ytest_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=1,\n",
        "        class_mode='binary')\n",
        "\n",
        "pred = VGG16_model.predict_generator(ytest_generator, verbose=1)\n",
        "\n",
        "predicted_class_indices = np.argmax(pred, axis=1)\n",
        "labels = (train_generator.class_indices)\n",
        "label = dict((v,k) for k,v in labels.items())\n",
        "\n",
        "# 建立代码标签与真实标签的关系\n",
        "predictions = [label[i] for i in predicted_class_indices]\n",
        "predictions\n",
        "#VGG16_model.predict_classes(ytest_generator, batch_size=len(ytest_generator), verbose=0)\n",
        "#ytest_loss, ytest_acc = VGG16_model.evaluate_generator(ytest_generator, steps=50)\n",
        "#print('ytest acc:', ytest_acc)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 2 images belonging to 2 classes.\n",
            "2/2 [==============================] - 0s 81ms/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['clean', 'clean']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vOknC3xvV4ni",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        },
        "outputId": "701621ec-ade3-4e51-ebb9-54d056419ce0"
      },
      "source": [
        "![picture](https://drive.google.com/file/d/1q4jo2YXhk-f168_eAJt2kyjbSCsStNv6/view?usp=sharing)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/bin/bash: -c: line 0: syntax error near unexpected token `https://drive.google.com/file/d/1q4jo2YXhk-f168_eAJt2kyjbSCsStNv6/view?usp=sharing'\n",
            "/bin/bash: -c: line 0: `[picture](https://drive.google.com/file/d/1q4jo2YXhk-f168_eAJt2kyjbSCsStNv6/view?usp=sharing)'\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XttJhgatWTZI",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "predict_dir = 'messy_vs_clean_room/predict'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jFjP-d9cWhLH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        },
        "outputId": "acf3242c-6d4b-4311-9ea6-f0172755c665"
      },
      "source": [
        "ytest_generator = ytest_datagen.flow_from_directory(\n",
        "        ytest_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=1,\n",
        "        class_mode='binary')\n",
        "\n",
        "test_loss, test_acc = VGG16_model.evaluate_generator(ytest_generator, steps=50)\n",
        "print('test acc:', test_acc)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 3 images belonging to 3 classes.\n",
            "test acc: 0.0\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5LJBaRyDYANZ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "7457ba77-430c-464d-cb7f-b6610c2f5b67"
      },
      "source": [
        "ls /content/messy_vs_clean_room/ytest"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\u001b[0m\u001b[01;34mClean\u001b[0m/  \u001b[01;34mMessy\u001b[0m/\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZXznXiUVW3RW",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "2531e24f-1112-48cf-c0a5-1c50f2a046fa"
      },
      "source": [
        "from PIL import Image\n",
        "import numpy as np\n",
        "from skimage import transform\n",
        "def load(filename):\n",
        "   np_image = Image.open(filename)\n",
        "   np_image = np.array(np_image).astype('float32')/255\n",
        "   np_image = transform.resize(np_image, (150, 150, 3))\n",
        "   np_image = np.expand_dims(np_image, axis=0)\n",
        "   return np_image\n",
        "image = load('/content/messy_vs_clean_room/ytest/Messy/Antes-y-después-de-cuartos-sucios-8.jpg')\n",
        "VGG16_model.predict_classes(image)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[1]], dtype=int32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 66
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rxUCbqozYjya",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "58a1a267-1887-445e-8f4f-896990eed8ce"
      },
      "source": [
        "import numpy as np\n",
        "predictions = VGG16_model.predict_generator(ytest_generator)\n",
        "y_pred = np.array([np.argmax(x) for x in predictions])\n",
        "y_pred.astype('int32')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0, 0, 0], dtype=int32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OjsXE_Z3Y55I",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 85
        },
        "outputId": "751f5b61-0136-4f37-9914-576f74caa83b"
      },
      "source": [
        "# Note that the validation data should not be augmented!\n",
        "predict_datagen = ImageDataGenerator(rescale=1./255)\n",
        "predict_generator = predict_datagen.flow_from_directory(\n",
        "        predict_dir,\n",
        "        target_size=(150, 150),\n",
        "        batch_size=2,\n",
        "        class_mode='binary')\n",
        "predict_loss, predict_acc = VGG16_model.evaluate_generator(predict_generator, steps=1)\n",
        "print('predict acc:', predict_acc)\n",
        "predictions = VGG16_model.predict_generator(predict_generator)\n",
        "predictions"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 2 images belonging to 2 classes.\n",
            "predict acc: 1.0\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[8.8349014e-04],\n",
              "       [9.9979502e-01]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 80
        }
      ]
    }
  ]
}