Deep-Sets-soln.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "similar-meditation",
   "metadata": {},
   "source": [
    "# Deep Sets Tutorial\n",
    "\n",
    "We were motivated to try Deep Sets ATLAS FTAG from the lovely \"Energy Flow Networks: Deep Sets for Particle Jets\" [paper](https://arxiv.org/abs/1810.05165) which includes a really nice pip installable [energyflow](https://energyflow.network/installation/) package as well.\n",
    "\n",
    "To end, we're using the same notation and variable names as the paper and code base, but this tutorial works with networks built in keras so we can do a few more studies with the network internals.\n",
    "\n",
    "Below are the topics that this tutorial covers:\n",
    "\n",
    "**Table of Contents:**\n",
    "1. [Loading in the dataset](#dataset)\n",
    "2. [Set up + train an architecture](#model)\n",
    "3. [Other permutation invariant operators](#permOps)\n",
    "4. [Including batch norm](#bn)\n",
    "\n",
    "Section (2) constitutes the bulk of the goals of this tutorial, while sections (3) and (4) give follow-up exercises to build our understanding by modifying different aspects of the network structure."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "cleared-evanescence",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "import h5py"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "incoming-accreditation",
   "metadata": {},
   "source": [
    "**Step 1:** Load in a preprocessed dataset\n",
    "<a name=\"dataset\"></a>\n",
    "\n",
    "The preprocessing is inside of the `Ntuple-Exploration notebook`.\n",
    "The (standard) dips inputs from [ATL-PHYS-PUB-2020-014](http://cds.cern.ch/record/2718948/files/ATL-PHYS-PUB-2020-014.pdf) are shown in the table below.\n",
    "<img src=\"dips-inputs.png\"\n",
    "     width=500/>\n",
    "\n",
    "\n",
    "Not all of these are present inside of the Ntuple right now, so we're working with an 8-dim input vector: \n",
    "`['sd0','sz0','ptfrac','dr','numPix', 'numSCT','d0','z0']`\n",
    "\n",
    "Additionally, jets that have less than 40 tracks are \"padded\" with zeros pass a fixed dimensional input vectors, although these padded tracks will be ignored in the computation.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "after-exception",
   "metadata": {},
   "outputs": [],
   "source": [
    "fDir = 'data' \n",
    "dname = f'{fDir}/output.hdf5'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "cubic-playing",
   "metadata": {},
   "outputs": [],
   "source": [
    "f = h5py.File(dname,\"r\")\n",
    "\n",
    "X_train       = f['X_train'][:]\n",
    "y_train       = f['y_train'][:]\n",
    "ix_train      = f['ix_train'][:]\n",
    "weights_train = f['weights_train'][:] # Weights to avoid learning directly from the pT dist\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "96893707",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "42a57d4a",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(array([ 0,  1,  2,  3, 44, 54, 55]),\n",
       " array([534,  59, 362,  38,   3,   3,   1]))"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.unique(y_train,return_counts=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d76bc7ed-f624-46a9-8884-d0fb46230e6c",
   "metadata": {},
   "source": [
    "**Category definitions:** The y vector is the truth label for the jet:\n",
    "- <span style=\"color:royalblue\"> 0: light-jet </span>. \n",
    "- <span style=\"color:orange\"> 1: c-jet </span>.\n",
    "- <span style=\"color:limegreen\"> 2: b-jet </span>\n",
    "- <span style=\"color:crimson\"> 3: $\\tau$-jet </span> - but we'll remove these from the training set\n",
    "\n",
    "<span style=\"color:royalblue\"> **Need to check these label definitions with Lorenzo!!** </span>. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "f6eb0ec2-2632-4d1d-8391-384f168539d4",
   "metadata": {},
   "outputs": [],
   "source": [
    "nClasses = 3 # Only consider l, c, and b-jets\n",
    "\n",
    "if y_train.max() >= nClasses:\n",
    "\n",
    "    valid_targets = (y_train < nClasses)\n",
    "\n",
    "    X_train       = X_train[valid_targets]\n",
    "    y_train       = y_train[valid_targets]\n",
    "    ix_train      = ix_train[valid_targets]\n",
    "    weights_train = weights_train[valid_targets]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2c7e93df-7e44-402e-839d-c476fbb14d04",
   "metadata": {},
   "source": [
    "**Transform the input vector to a one-hot vector**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "be1a229e-75ef-4212-af13-44128a7aa325",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([2, 0, 2, 0])"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train[:4]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "4eed0984-e468-4a1e-8911-6004d2431fa8",
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.utils.np_utils import to_categorical\n",
    "\n",
    "y_train_cat = to_categorical(y_train, num_classes=nClasses)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "8ca28d11-34de-467b-bcde-b6d9c0f18117",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[0., 0., 1.],\n",
       "       [1., 0., 0.],\n",
       "       [0., 0., 1.],\n",
       "       [1., 0., 0.]], dtype=float32)"
      ]
     },
     "execution_count": 70,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "y_train_cat[:4]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "particular-ghost",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(955, 40, 8)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "banner-radical",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "ready-heath",
   "metadata": {},
   "source": [
    "**Step 2:** Set up a Deep Sets architecture\n",
    "<a name=\"model\"></a>\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "43664b4c-9541-46d4-95ef-bdaab9ed28d5",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow import keras\n",
    "from keras.layers import BatchNormalization, Layer, TimeDistributed\n",
    "from keras.layers import Dense, Input, ReLU, Masking\n",
    "from keras.models import Model\n",
    "\n",
    "from keras.callbacks import EarlyStopping, ModelCheckpoint\n",
    "from keras import backend as K"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "defc2b1e-6422-4d0a-9567-8858041e29f0",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "solved-springer",
   "metadata": {},
   "source": [
    "Keras is very modular and lets you stack layers like legos to build a custom architecture. For the Deep-Sets model, we just needed a custom layer to implement this sum over tracks, which is implemented in the block below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "3337e6a8-0f8b-4535-afcf-fb4a51dd8820",
   "metadata": {},
   "outputs": [],
   "source": [
    "class Sum(Layer):\n",
    "    \"\"\"\n",
    "    Simple sum layer à la Dan Guest\n",
    "    \"\"\"\n",
    "\n",
    "    def __init__(self, **kwargs):\n",
    "        super().__init__(**kwargs)\n",
    "        self.supports_masking = True\n",
    "\n",
    "    def build(self, input_shape):\n",
    "        pass\n",
    "\n",
    "    def call(self, x, mask=None):\n",
    "        if mask is not None:\n",
    "            x = x * K.cast(mask, K.dtype(x))[:,:,None]\n",
    "        return K.sum(x, axis=1)\n",
    "\n",
    "    def compute_output_shape(self, input_shape):\n",
    "        return input_shape[0], input_shape[2]\n",
    "\n",
    "    def compute_mask(self, inputs, mask):\n",
    "        return None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "useful-preservation",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "handled-anniversary",
   "metadata": {},
   "source": [
    "**Set up the model skeleton** \n",
    "\n",
    "The model parameters are set as variables to be easy to change in the tutorial.\n",
    "\n",
    "- `ppm_sizes_int`: A list defining the size for each of the dense layers of the per particle network Φ\n",
    "- `dense_sizes_int`: A list defining the size for each of the dense layers of the per particle network F"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "visible-national",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Store the dimensions of the input dataset for setting up the model\n",
    "nJets, maxNumTrks, nFeatures = X_train.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "04171f16-cc36-42b7-9af9-356f4f0d8fb1",
   "metadata": {},
   "source": [
    "<img src=\"DIPS_architecture_tutorial.pdf\"\n",
    "     width=500\n",
    "     style=\"float: left; margin-right: 10px;\" />"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7b7b3c24-bda7-421f-b331-9bebcf04dc46",
   "metadata": {},
   "source": [
    "**TO DO:** Define the ppm_sizes_int and the dense_sizes_int vectors to implement the architecure above."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "bright-things",
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "TO DO: fill in this cell\n",
    "'''\n",
    "\n",
    "ppm_sizes_int = [50,50,32]\n",
    "dense_sizes_int = [50,50]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "distant-classroom",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-06-02 23:13:05.475432: W tensorflow/stream_executor/platform/default/dso_loader.cc:64] Could not load dynamic library 'libcuda.so.1'; dlerror: libcuda.so.1: cannot open shared object file: No such file or directory; LD_LIBRARY_PATH: /cvmfs/sft.cern.ch/lcg/releases/MCGenerators/thepeg/2.2.3-6cb32/x86_64-centos7-gcc11-opt/lib/ThePEG:/cvmfs/sft.cern.ch/lcg/releases/MCGenerators/herwig++/7.2.3-90caf/x86_64-centos7-gcc11-opt/lib/Herwig:/cvmfs/sft.cern.ch/lcg/views/LCG_102b_swan/x86_64-centos7-gcc11-opt/lib/python3.9/site-packages/jaxlib/mlir/_mlir_libs:/cvmfs/sft.cern.ch/lcg/views/LCG_102b_swan/x86_64-centos7-gcc11-opt/lib/python3.9/site-packages/torch/lib:/cvmfs/sft.cern.ch/lcg/views/LCG_102b_swan/x86_64-centos7-gcc11-opt/lib/python3.9/site-packages/onnxruntime/capi/:/cvmfs/sft.cern.ch/lcg/views/LCG_102b_swan/x86_64-centos7-gcc11-opt/lib/python3.9/site-packages/tensorflow:/cvmfs/sft.cern.ch/lcg/views/LCG_102b_swan/x86_64-centos7-gcc11-opt/lib/python3.9/site-packages/tensorflow/contrib/tensor_forest:/cvmfs/sft.cern.ch/lcg/views/LCG_102b_swan/x86_64-centos7-gcc11-opt/lib/python3.9/site-packages/tensorflow/python/framework:/cvmfs/sft.cern.ch/lcg/releases/java/11.0.14p1-8284a/x86_64-centos7-gcc11-opt/jre/lib/amd64:/cvmfs/sft.cern.ch/lcg/views/LCG_102b_swan/x86_64-centos7-gcc11-opt/lib64:/cvmfs/sft.cern.ch/lcg/views/LCG_102b_swan/x86_64-centos7-gcc11-opt/lib:/cvmfs/sft.cern.ch/lcg/releases/gcc/11.3.0-ad0f5/x86_64-centos7/lib:/cvmfs/sft.cern.ch/lcg/releases/gcc/11.3.0-ad0f5/x86_64-centos7/lib64:/cvmfs/sft.cern.ch/lcg/releases/binutils/2.37-355ed/x86_64-centos7/lib:/usr/local/lib/:/cvmfs/sft.cern.ch/lcg/releases/R/4.1.2-f1888/x86_64-centos7-gcc11-opt/lib64/R/library/readr/rcon\n",
      "2023-06-02 23:13:05.475504: W tensorflow/stream_executor/cuda/cuda_driver.cc:269] failed call to cuInit: UNKNOWN ERROR (303)\n",
      "2023-06-02 23:13:05.475724: I tensorflow/stream_executor/cuda/cuda_diagnostics.cc:163] no NVIDIA GPU device is present: /dev/nvidia0 does not exist\n",
      "2023-06-02 23:13:05.476697: I tensorflow/core/platform/cpu_feature_guard.cc:151] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations:  AVX2 AVX512F FMA\n",
      "To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    }
   ],
   "source": [
    "trk_inputs = Input(shape=(maxNumTrks,nFeatures),name='Input')\n",
    "masked_inputs = Masking(mask_value=0,name='Mask')(trk_inputs)\n",
    "tdd = masked_inputs\n",
    "\n",
    "for i, phi_nodes in enumerate(ppm_sizes_int):\n",
    "\n",
    "    tdd = TimeDistributed(Dense(phi_nodes,activation='linear'),name=f\"Phi{i}_Dense\")(tdd)\n",
    "    tdd = TimeDistributed(ReLU(),name=f\"Phi{i}_ReLU\")(tdd)\n",
    "\n",
    "# This is where the magic happens... sum up the track features!\n",
    "F = Sum(name=\"Sum\")(tdd)\n",
    "\n",
    "for j, F_nodes in enumerate(dense_sizes_int):\n",
    "\n",
    "    F = Dense(F_nodes, activation='linear', name=f\"F{j}_Dense\")(F)\n",
    "    F = ReLU(name=f\"F{j}_ReLU\")(F)\n",
    "\n",
    "output = Dense(nClasses, activation='softmax',name=\"Jet_class\")(F)\n",
    "dips = Model(inputs=trk_inputs, outputs=output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "ab3c927d-f4fa-4292-8389-4835c44fa808",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Input True\n",
      "Mask True\n",
      "Phi0_Dense True\n",
      "Phi0_ReLU True\n",
      "Phi1_Dense True\n",
      "Phi1_ReLU True\n",
      "Phi2_Dense True\n",
      "Phi2_ReLU True\n",
      "Sum True\n",
      "F0_Dense True\n",
      "F0_ReLU True\n",
      "F1_Dense True\n",
      "F1_ReLU True\n",
      "Jet_class True\n"
     ]
    }
   ],
   "source": [
    "for l in dips.layers:\n",
    "    print(l.name, l.supports_masking)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "removed-logistics",
   "metadata": {},
   "source": [
    "As a sanitry check, you can check the structure of the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "hairy-racing",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " Input (InputLayer)          [(None, 40, 8)]           0         \n",
      "                                                                 \n",
      " Mask (Masking)              (None, 40, 8)             0         \n",
      "                                                                 \n",
      " Phi0_Dense (TimeDistributed  (None, 40, 50)           450       \n",
      " )                                                               \n",
      "                                                                 \n",
      " Phi0_ReLU (TimeDistributed)  (None, 40, 50)           0         \n",
      "                                                                 \n",
      " Phi1_Dense (TimeDistributed  (None, 40, 50)           2550      \n",
      " )                                                               \n",
      "                                                                 \n",
      " Phi1_ReLU (TimeDistributed)  (None, 40, 50)           0         \n",
      "                                                                 \n",
      " Phi2_Dense (TimeDistributed  (None, 40, 32)           1632      \n",
      " )                                                               \n",
      "                                                                 \n",
      " Phi2_ReLU (TimeDistributed)  (None, 40, 32)           0         \n",
      "                                                                 \n",
      " Sum (Sum)                   (None, 32)                0         \n",
      "                                                                 \n",
      " F0_Dense (Dense)            (None, 50)                1650      \n",
      "                                                                 \n",
      " F0_ReLU (ReLU)              (None, 50)                0         \n",
      "                                                                 \n",
      " F1_Dense (Dense)            (None, 50)                2550      \n",
      "                                                                 \n",
      " F1_ReLU (ReLU)              (None, 50)                0         \n",
      "                                                                 \n",
      " Jet_class (Dense)           (None, 3)                 153       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 8,985\n",
      "Trainable params: 8,985\n",
      "Non-trainable params: 0\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "dips.summary() "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "opened-source",
   "metadata": {},
   "source": [
    "To train the model, you need to define what you mean by a model performing better or worse, which we quantify by a loss function. For multi-class classification, we use the categorical cross-entropy, which maximizes the node corresponding to the true label of the jet.You also need a strategy for updating the weights as you minimize the loss, which is inside the optimizer. There are many different optimization strategies on the market and available in Keras, but the optimizer adam is often a good one to start with since it provides a way to decrease the weight update as you converge onto a solution.This is what is done in the \"compile\" function below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "everyday-absence",
   "metadata": {},
   "outputs": [],
   "source": [
    "opt = keras.optimizers.Adam(learning_rate=0.0001)\n",
    "\n",
    "dips.compile(loss='categorical_crossentropy', optimizer='adam',\n",
    "             metrics=['acc'])  "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "assigned-municipality",
   "metadata": {},
   "source": [
    "**Warm up Q:** _Before training_ the model, what do you think the accuracy will be?\n",
    "\n",
    "\n",
    "- A.   0% \n",
    "- B.   33%\n",
    "- C.   50%\n",
    "- D.   100%"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "liable-thousand",
   "metadata": {},
   "source": [
    "**Test your answer above:** "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "welsh-village",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "8/8 [==============================] - 1s 6ms/step - loss: 5222.1973 - acc: 0.3791\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[5222.197265625, 0.3790575861930847]"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dips.evaluate(X_train,y_train_cat,batch_size=128)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "superior-shore",
   "metadata": {},
   "source": [
    "The second number above is the accuracy, which is close to 1/3, since before training the model is just guessing among the three outputs.The first number is the loss, which gets minimized in the training."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "passive-yacht",
   "metadata": {},
   "source": [
    "**Training details**\n",
    "\n",
    "When training a model, how do you decide how long to go for? When training NNs, you look at your training dataset multiple times as you converge to a solution. A single pass over the training dataset is called an epoch.\n",
    "\n",
    "The model nEpochs is set at 7 right now and is a variable for the maximimum number of epochs, but in practice, it's common to set this value to something larger, and use the loss on a separate validation dataset to decide when to stop training. This is set up in the EarlyStopping callback which says to stop training when the loss on the validation dataset (val_loss) has not improved in 10 epochs.\n",
    "\n",
    "Then the ModelCheckpoint below saves the name for the model weights file you're using."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "ahead-clerk",
   "metadata": {},
   "outputs": [],
   "source": [
    "nEpochs = 100\n",
    "\n",
    "earlyStop = EarlyStopping(monitor='val_loss', verbose=True, patience=10)\n",
    "\n",
    "dips_mChkPt = ModelCheckpoint('dips_weights.h5',\n",
    "                              monitor='val_loss', \n",
    "                              verbose=True,\n",
    "                              save_best_only=True,\n",
    "                              save_weights_only=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "trying-billy",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "1/6 [====>.........................] - ETA: 6s - loss: 5575.2500 - acc: 0.3828\n",
      "Epoch 1: val_loss improved from inf to 332.28506, saving model to dips_weights.h5\n",
      "6/6 [==============================] - 2s 69ms/step - loss: 2340.1099 - acc: 0.4280 - val_loss: 332.2851 - val_acc: 0.4817\n",
      "Epoch 2/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 168.1262 - acc: 0.5625\n",
      "Epoch 2: val_loss improved from 332.28506 to 327.31787, saving model to dips_weights.h5\n",
      "6/6 [==============================] - 0s 24ms/step - loss: 644.4141 - acc: 0.4725 - val_loss: 327.3179 - val_acc: 0.5550\n",
      "Epoch 3/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 553.6637 - acc: 0.5703\n",
      "Epoch 3: val_loss improved from 327.31787 to 192.15919, saving model to dips_weights.h5\n",
      "6/6 [==============================] - 0s 23ms/step - loss: 332.5081 - acc: 0.5131 - val_loss: 192.1592 - val_acc: 0.4398\n",
      "Epoch 4/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 235.4822 - acc: 0.5078\n",
      "Epoch 4: val_loss improved from 192.15919 to 150.05156, saving model to dips_weights.h5\n",
      "6/6 [==============================] - 0s 29ms/step - loss: 212.4712 - acc: 0.4411 - val_loss: 150.0516 - val_acc: 0.5445\n",
      "Epoch 5/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 168.7674 - acc: 0.5781\n",
      "Epoch 5: val_loss improved from 150.05156 to 91.67184, saving model to dips_weights.h5\n",
      "6/6 [==============================] - 0s 23ms/step - loss: 243.6502 - acc: 0.5157 - val_loss: 91.6718 - val_acc: 0.5026\n",
      "Epoch 6/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 199.9493 - acc: 0.5156\n",
      "Epoch 6: val_loss did not improve from 91.67184\n",
      "6/6 [==============================] - 0s 15ms/step - loss: 155.4184 - acc: 0.4372 - val_loss: 99.7916 - val_acc: 0.4450\n",
      "Epoch 7/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 124.3754 - acc: 0.5234\n",
      "Epoch 7: val_loss did not improve from 91.67184\n",
      "6/6 [==============================] - 0s 14ms/step - loss: 162.4772 - acc: 0.4319 - val_loss: 259.7905 - val_acc: 0.4764\n",
      "Epoch 8/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 204.5101 - acc: 0.5547\n",
      "Epoch 8: val_loss improved from 91.67184 to 88.62811, saving model to dips_weights.h5\n",
      "6/6 [==============================] - 0s 23ms/step - loss: 231.9458 - acc: 0.4869 - val_loss: 88.6281 - val_acc: 0.4660\n",
      "Epoch 9/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 235.6864 - acc: 0.4062\n",
      "Epoch 9: val_loss did not improve from 88.62811\n",
      "6/6 [==============================] - 0s 19ms/step - loss: 243.9453 - acc: 0.5249 - val_loss: 145.5356 - val_acc: 0.4293\n",
      "Epoch 10/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 227.0353 - acc: 0.3984\n",
      "Epoch 10: val_loss did not improve from 88.62811\n",
      "6/6 [==============================] - 0s 15ms/step - loss: 242.0436 - acc: 0.4215 - val_loss: 337.4002 - val_acc: 0.5026\n",
      "Epoch 11/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 406.2858 - acc: 0.5469\n",
      "Epoch 11: val_loss did not improve from 88.62811\n",
      "6/6 [==============================] - 0s 15ms/step - loss: 292.9503 - acc: 0.4568 - val_loss: 219.3537 - val_acc: 0.4660\n",
      "Epoch 12/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 214.8867 - acc: 0.5234\n",
      "Epoch 12: val_loss did not improve from 88.62811\n",
      "6/6 [==============================] - 0s 14ms/step - loss: 218.2969 - acc: 0.4921 - val_loss: 193.3909 - val_acc: 0.4241\n",
      "Epoch 13/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 240.0671 - acc: 0.4297\n",
      "Epoch 13: val_loss did not improve from 88.62811\n",
      "6/6 [==============================] - 0s 15ms/step - loss: 201.7641 - acc: 0.4725 - val_loss: 100.9665 - val_acc: 0.4293\n",
      "Epoch 14/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 153.8873 - acc: 0.4375\n",
      "Epoch 14: val_loss did not improve from 88.62811\n",
      "6/6 [==============================] - 0s 16ms/step - loss: 136.7867 - acc: 0.5393 - val_loss: 127.7225 - val_acc: 0.4607\n",
      "Epoch 15/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 199.1978 - acc: 0.3828\n",
      "Epoch 15: val_loss did not improve from 88.62811\n",
      "6/6 [==============================] - 0s 15ms/step - loss: 142.2140 - acc: 0.4555 - val_loss: 167.7478 - val_acc: 0.5864\n",
      "Epoch 16/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 213.3399 - acc: 0.5781\n",
      "Epoch 16: val_loss did not improve from 88.62811\n",
      "6/6 [==============================] - 0s 16ms/step - loss: 143.0016 - acc: 0.5183 - val_loss: 100.8190 - val_acc: 0.4031\n",
      "Epoch 17/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 111.4728 - acc: 0.3281\n",
      "Epoch 17: val_loss did not improve from 88.62811\n",
      "6/6 [==============================] - 0s 14ms/step - loss: 154.2820 - acc: 0.4581 - val_loss: 123.6056 - val_acc: 0.5340\n",
      "Epoch 18/100\n",
      "1/6 [====>.........................] - ETA: 0s - loss: 170.2368 - acc: 0.5547\n",
      "Epoch 18: val_loss did not improve from 88.62811\n",
      "6/6 [==============================] - 0s 14ms/step - loss: 121.1765 - acc: 0.4660 - val_loss: 100.2090 - val_acc: 0.5707\n",
      "Epoch 18: early stopping\n"
     ]
    }
   ],
   "source": [
    "dips_hist = dips.fit(X_train, y_train_cat, epochs=nEpochs, \n",
    "                     batch_size=128,validation_split=0.2,\n",
    "                     callbacks=[earlyStop, dips_mChkPt])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "06d773b3-b575-4cef-8e22-b977dfd7d27c",
   "metadata": {},
   "source": [
    "Let's look at the training curve!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "obvious-seafood",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "epochs = np.arange(1,len(dips_hist.history['loss'])+1)\n",
    "\n",
    "plt.plot(epochs,dips_hist.history['loss'],color='hotpink',label='training')\n",
    "plt.plot(epochs,dips_hist.history['val_loss'],color='hotpink',label='validation',ls='--')\n",
    "plt.xlabel('epochs',fontsize=14)\n",
    "plt.ylabel('cross-entropy loss',fontsize=14)\n",
    "plt.legend()\n",
    "plt.title('DIPS')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3c9a8ee0-d172-42fb-a3a1-122da2a5304d",
   "metadata": {},
   "source": [
    "Since only 1k jets are being used in the training rn, the model might be overfitting for these hyperparameters."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "commercial-microphone",
   "metadata": {},
   "source": [
    "**Evaluate the results with a roc curve**\n",
    "\n",
    "For this multiclass output problem, we'll combine the three class probabilities into a single discriminant.\n",
    "\n",
    "$$D_b = \\log \\frac{p_b}{f_c \\cdot p_c + (1 - f_c) \\cdot p_l}$$\n",
    "\n",
    "where $f_c$ is the \"charm fraction\" which is a parameter we choose after training the network.\n",
    "\n",
    "For this tutorial we will set $f_c = 0.07$.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "07e56335-4585-46e5-8576-addbbade9f0a",
   "metadata": {},
   "source": [
    "**(Warm up) question:** Will a $b$-jet have high or low values of $D_b$?\n",
    "\n",
    "-> High values of $D_b$\n",
    "\n",
    "What about $c$ and $l$-jets?\n",
    "\n",
    "-> Low values of $D_b$"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "39505fc3-181b-47f5-b2f7-97ed6e82a9ec",
   "metadata": {},
   "source": [
    "**Test your assumption** The code below does the roc curve calculation.\n",
    "\n",
    "TO DO: Fill in the discriminant calculation (the line below starting with)\n",
    "\n",
    "`disc = ...` "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "id": "8be2cab1-4ff2-4424-aa39-03ce5148bd94",
   "metadata": {},
   "outputs": [],
   "source": [
    "def sigBkgEff(myModel, X_test, y_test, fc=0.07, title=''):\n",
    "    '''\n",
    "    Given a model, make the histograms of the model outputs to get the ROC curves.\n",
    "\n",
    "    Input:\n",
    "        myModel: A keras model\n",
    "        X_test: Model inputs of the test set\n",
    "        y_test: Truth labels for the test set\n",
    "        fc: The amount by which to weight the c-jet prob in the disc. The\n",
    "            default value of 0.07 corresponds to the fraction of c-jet bkg\n",
    "            in ttbar.\n",
    "\n",
    "    Output:\n",
    "        effs: A list with 3 entries for the l, c, and b effs\n",
    "    '''\n",
    "\n",
    "    # Evaluate the performance with the ROC curves!\n",
    "    predictions = myModel.predict(X_test,verbose=True)\n",
    "\n",
    "    '''\n",
    "    TO DO: Set up the discriminant function\n",
    "    '''\n",
    "    disc = np.log(np.divide(predictions[:,2], fc*predictions[:,1] + (1 - fc) * predictions[:,0]))\n",
    "    \n",
    "    # Define the min and max range for the distribution\n",
    "    real = (disc != -np.inf) & (disc != np.inf)\n",
    "    np.min(disc[real]),np.max(disc[real])\n",
    "    \n",
    "    discMin = np.min(disc[real])\n",
    "    discMax = np.max(disc[real])\n",
    "    print(\"Db discriminant range:\",discMin,discMax)\n",
    "    \n",
    "    # Reset the -inf and +inf to have ranges at the edge of the distribution\n",
    "    eps=1e-10\n",
    "    disc[disc == -np.inf] = discMin+eps\n",
    "    disc[disc ==  np.inf] = discMax-eps\n",
    "   \n",
    "    myRange=(discMin,discMax)\n",
    "    nBins = 200\n",
    "\n",
    "    effs = []\n",
    "    for i, flavor in zip([0,1,2], ['l','c','b']):\n",
    "\n",
    "        ix = (y_test == i)\n",
    "        \n",
    "        # Plot the discriminant output\n",
    "        # nEntries is just a sum of the weight of each bin in the histogram.\n",
    "        hi = plt.hist(disc[ix],nBins, myRange, alpha=0.5, density=True,\n",
    "                      label=f'{flavor}-jets', log=True)[0]\n",
    "        \n",
    "        # Since high Db scores correspond to more b-like jets, compute the cummulative density function\n",
    "        # from summing from high to low values, this is why we reverse the order of the bins in nEntries\n",
    "        # using the \"::-1\" numpy indexing.\n",
    "        eff = np.add.accumulate(hi[::-1]) / np.sum(hi)\n",
    "        effs.append(eff)\n",
    "\n",
    "    plt.title(title)\n",
    "    plt.legend()\n",
    "    plt.xlabel('$D = \\ln [ p_b / (f_c p_c + (1- f_c)p_l ) ]$',fontsize=14)\n",
    "    plt.ylabel('Normalized entries')\n",
    "        \n",
    "    return effs\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "986a143f-f8b5-4a15-b361-0fc30dd1bad4",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_test = f['X_test'][:]\n",
    "y_test = f['y_test'][:]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "226b2190",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "organic-bibliography",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "79/79 [==============================] - 0s 3ms/step\n",
      "Db discriminant range: -87.110535 86.89606\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_7268/2548379725.py:23: RuntimeWarning: divide by zero encountered in true_divide\n",
      "  disc = np.log(np.divide(predictions[:,2], fc*predictions[:,1] + (1 - fc) * predictions[:,0]))\n",
      "/tmp/ipykernel_7268/2548379725.py:23: RuntimeWarning: divide by zero encountered in log\n",
      "  disc = np.log(np.divide(predictions[:,2], fc*predictions[:,1] + (1 - fc) * predictions[:,0]))\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "title = 'DIPS: $\\Phi$ ' + '-'.join([str(i) for i in ppm_sizes_int])\n",
    "title += ', F ' + '-'.join([str(i) for i in dense_sizes_int])\n",
    "\n",
    "leff, ceff, beff = sigBkgEff(dips, X_test, y_test, title=title)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "emerging-ivory",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# l-rej\n",
    "plt.figure()\n",
    "plt.plot(beff, 1 / leff, color='hotpink', label='dips: l-rej')\n",
    "\n",
    "# c-rej\n",
    "plt.plot(beff, 1 / ceff, color='hotpink', linestyle='--', label='dips: c-rej')\n",
    "plt.xlabel('b efficiency')\n",
    "plt.ylabel('Background rejection')\n",
    "\n",
    "plt.legend()\n",
    "plt.yscale(\"log\")\n",
    "plt.xlim(0.6,1)\n",
    "# plt.ylim(0,int(2e3))\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "accessory-bracelet",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "fifteen-temperature",
   "metadata": {},
   "source": [
    "**Step 3:** Test out other permutation invariant operations\n",
    "<a name=\"permOps\"></a>\n",
    "\n",
    "- max\n",
    "- mean\n",
    "\n",
    "**Hint(s):** \n",
    "- Use the Sum() layer above for inspiration\n",
    "- Try typing K. and press the tab completion to see what other functions are available"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "secondary-emergency",
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "Will be left blank for the students\n",
    "'''\n",
    "\n",
    "class Max(Layer):\n",
    "    def __init__(self, **kwargs):\n",
    "        super().__init__(**kwargs)\n",
    "        self.supports_masking = True\n",
    "\n",
    "    def build(self, input_shape):\n",
    "        pass\n",
    "\n",
    "    def call(self, x, mask=None):\n",
    "        if mask is not None:\n",
    "            x = x * K.cast(mask, K.dtype(x))[:,:,None]\n",
    "        return K.max(x, axis=1)\n",
    "\n",
    "    def compute_output_shape(self, input_shape):\n",
    "        return input_shape[0], input_shape[2]\n",
    "\n",
    "    def compute_mask(self, inputs, mask):\n",
    "        return None"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "round-display",
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "Will be left blank for the students\n",
    "'''\n",
    "\n",
    "class Mean(Layer):\n",
    "    def __init__(self, **kwargs):\n",
    "        super().__init__(**kwargs)\n",
    "        self.supports_masking = True\n",
    "\n",
    "    def build(self, input_shape):\n",
    "        pass\n",
    "\n",
    "    def call(self, x, mask=None):\n",
    "        if mask is not None:\n",
    "            x = x * K.cast(mask, K.dtype(x))[:,:,None]\n",
    "        return K.mean(x, axis=1)\n",
    "\n",
    "    def compute_output_shape(self, input_shape):\n",
    "        return input_shape[0], input_shape[2]\n",
    "\n",
    "    def compute_mask(self, inputs, mask):\n",
    "        return None"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "bbf5dcc3-34b6-46ff-b36b-cf2b437bbe97",
   "metadata": {},
   "source": [
    "**Try training these models and see how they compare:**\n",
    "\n",
    "It's fine to use the same hyperparameters as above for the $\\Phi$ and $F$ networks to just compare the impact of the pooling operation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "premier-administration",
   "metadata": {},
   "outputs": [],
   "source": [
    "trk_inputs = Input(shape=(maxNumTrks,nFeatures),name='Input')\n",
    "masked_inputs = Masking(mask_value=0)(trk_inputs)\n",
    "tdd = masked_inputs\n",
    "\n",
    "for i, phi_nodes in enumerate(ppm_sizes_int):\n",
    "\n",
    "    tdd = TimeDistributed(Dense(phi_nodes,activation='linear'),name=f\"Phi{i}_Dense\")(tdd)\n",
    "    tdd = TimeDistributed(ReLU(),name=f\"Phi{i}_ReLU\")(tdd)\n",
    "\n",
    "# This is where the magic happens... sum up the track features!\n",
    "F = Max(name=\"Max\")(tdd)\n",
    "\n",
    "for j, F_nodes in enumerate(dense_sizes_int):\n",
    "\n",
    "    F = Dense(F_nodes, activation='linear', name=f\"F{j}_Dense\")(F)\n",
    "    F = ReLU(name=f\"F{j}_ReLU\")(F)\n",
    "\n",
    "output = Dense(nClasses, activation='softmax',name=\"Jet_class\")(F)\n",
    "\n",
    "'''\n",
    "All but the last line here will be left blank for the students\n",
    "'''\n",
    "\n",
    "dips_max = Model(inputs=trk_inputs, outputs=output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "d78ceed0-757a-4f7d-a489-db6cb3575038",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 2698.8303 - acc: 0.3212\n",
      "Epoch 1: val_loss did not improve from 6.86492\n",
      "30/30 [==============================] - 2s 21ms/step - loss: 2564.6482 - acc: 0.3226 - val_loss: 995.9360 - val_acc: 0.3312\n",
      "Epoch 2/100\n",
      "25/30 [========================>.....] - ETA: 0s - loss: 387.3573 - acc: 0.3519\n",
      "Epoch 2: val_loss did not improve from 6.86492\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 343.4956 - acc: 0.3875 - val_loss: 91.4044 - val_acc: 0.5791\n",
      "Epoch 3/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 46.4082 - acc: 0.4872\n",
      "Epoch 3: val_loss did not improve from 6.86492\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 44.7508 - acc: 0.4875 - val_loss: 16.3094 - val_acc: 0.5770\n",
      "Epoch 4/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 14.0889 - acc: 0.4688\n",
      "Epoch 4: val_loss improved from 6.86492 to 5.76495, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 14ms/step - loss: 13.3797 - acc: 0.4743 - val_loss: 5.7650 - val_acc: 0.3671\n",
      "Epoch 5/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 6.1904 - acc: 0.4962\n",
      "Epoch 5: val_loss improved from 5.76495 to 4.96752, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 13ms/step - loss: 6.2209 - acc: 0.4898 - val_loss: 4.9675 - val_acc: 0.5580\n",
      "Epoch 6/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 6.1061 - acc: 0.4958\n",
      "Epoch 6: val_loss did not improve from 4.96752\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 5.9661 - acc: 0.4983 - val_loss: 5.9720 - val_acc: 0.3460\n",
      "Epoch 7/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 7.5446 - acc: 0.4902\n",
      "Epoch 7: val_loss improved from 4.96752 to 4.78543, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 12ms/step - loss: 7.7552 - acc: 0.4988 - val_loss: 4.7854 - val_acc: 0.3534\n",
      "Epoch 8/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 7.3633 - acc: 0.5061\n",
      "Epoch 8: val_loss did not improve from 4.78543\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 7.2040 - acc: 0.5044 - val_loss: 6.6657 - val_acc: 0.6255\n",
      "Epoch 9/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 5.5144 - acc: 0.4951\n",
      "Epoch 9: val_loss did not improve from 4.78543\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 5.9610 - acc: 0.5009 - val_loss: 7.7108 - val_acc: 0.3914\n",
      "Epoch 10/100\n",
      "26/30 [=========================>....] - ETA: 0s - loss: 6.8795 - acc: 0.4973\n",
      "Epoch 10: val_loss improved from 4.78543 to 2.95447, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 12ms/step - loss: 6.5343 - acc: 0.5004 - val_loss: 2.9545 - val_acc: 0.3998\n",
      "Epoch 11/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 4.4167 - acc: 0.5204\n",
      "Epoch 11: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 4.4301 - acc: 0.5239 - val_loss: 3.8022 - val_acc: 0.5580\n",
      "Epoch 12/100\n",
      "26/30 [=========================>....] - ETA: 0s - loss: 4.2050 - acc: 0.5219\n",
      "Epoch 12: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 4.4642 - acc: 0.5276 - val_loss: 8.0904 - val_acc: 0.3787\n",
      "Epoch 13/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 7.9882 - acc: 0.4994\n",
      "Epoch 13: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 7.8695 - acc: 0.5033 - val_loss: 5.0022 - val_acc: 0.5928\n",
      "Epoch 14/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 7.1409 - acc: 0.5128\n",
      "Epoch 14: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 7.1488 - acc: 0.5104 - val_loss: 13.4824 - val_acc: 0.2542\n",
      "Epoch 15/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 5.8582 - acc: 0.5213\n",
      "Epoch 15: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 5.7785 - acc: 0.5239 - val_loss: 3.5993 - val_acc: 0.5886\n",
      "Epoch 16/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 4.4855 - acc: 0.5594\n",
      "Epoch 16: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 4.6904 - acc: 0.5603 - val_loss: 3.4676 - val_acc: 0.5232\n",
      "Epoch 17/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 4.4307 - acc: 0.5345\n",
      "Epoch 17: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 4.4969 - acc: 0.5352 - val_loss: 3.7947 - val_acc: 0.6013\n",
      "Epoch 18/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 4.8383 - acc: 0.5426\n",
      "Epoch 18: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 4.8426 - acc: 0.5460 - val_loss: 7.1006 - val_acc: 0.6245\n",
      "Epoch 19/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 8.2505 - acc: 0.5318\n",
      "Epoch 19: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 8.3654 - acc: 0.5379 - val_loss: 6.2869 - val_acc: 0.4494\n",
      "Epoch 20/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 6.2367 - acc: 0.5075\n",
      "Epoch 20: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 6.3142 - acc: 0.5194 - val_loss: 3.6027 - val_acc: 0.5675\n",
      "Epoch 20: early stopping\n"
     ]
    }
   ],
   "source": [
    "opt = keras.optimizers.Adam(learning_rate=0.0001)\n",
    "\n",
    "dips_max.compile(loss='categorical_crossentropy', optimizer=opt,\n",
    "                 metrics=['acc'])  \n",
    "\n",
    "dips_hist_max = dips_max.fit(X_train, y_train_cat, epochs=nEpochs, \n",
    "                         batch_size=128,validation_split=0.2,\n",
    "                         callbacks=[earlyStop, dips_mChkPt])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "648b09e3-9571-4a28-becb-ecd5fbe7826d",
   "metadata": {},
   "outputs": [],
   "source": [
    "trk_inputs = Input(shape=(maxNumTrks,nFeatures),name='Input')\n",
    "masked_inputs = Masking(mask_value=0)(trk_inputs)\n",
    "tdd = masked_inputs\n",
    "\n",
    "for i, phi_nodes in enumerate(ppm_sizes_int):\n",
    "\n",
    "    tdd = TimeDistributed(Dense(phi_nodes,activation='linear'),name=f\"Phi{i}_Dense\")(tdd)\n",
    "    tdd = TimeDistributed(ReLU(),name=f\"Phi{i}_ReLU\")(tdd)\n",
    "\n",
    "# This is where the magic happens... sum up the track features!\n",
    "F = Mean(name=\"Mean\")(tdd)\n",
    "\n",
    "for j, F_nodes in enumerate(dense_sizes_int):\n",
    "\n",
    "    F = Dense(F_nodes, activation='linear', name=f\"F{j}_Dense\")(F)\n",
    "    F = ReLU(name=f\"F{j}_ReLU\")(F)\n",
    "\n",
    "output = Dense(nClasses, activation='softmax',name=\"Jet_class\")(F)\n",
    "\n",
    "'''\n",
    "All but the last line here will be left blank for the students\n",
    "'''\n",
    "\n",
    "dips_mean = Model(inputs=trk_inputs, outputs=output)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "15438184-26b8-4ad7-927f-c14a7ef5b861",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 269.5942 - acc: 0.3323\n",
      "Epoch 1: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 2s 21ms/step - loss: 264.6492 - acc: 0.3347 - val_loss: 140.9984 - val_acc: 0.3892\n",
      "Epoch 2/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 107.8596 - acc: 0.3900\n",
      "Epoch 2: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 101.4134 - acc: 0.3980 - val_loss: 27.6659 - val_acc: 0.4863\n",
      "Epoch 3/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 19.7468 - acc: 0.4913\n",
      "Epoch 3: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 19.5649 - acc: 0.4885 - val_loss: 9.3093 - val_acc: 0.4810\n",
      "Epoch 4/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 10.4356 - acc: 0.4884\n",
      "Epoch 4: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 9.9076 - acc: 0.4912 - val_loss: 5.0060 - val_acc: 0.5190\n",
      "Epoch 5/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 5.9069 - acc: 0.5084\n",
      "Epoch 5: val_loss did not improve from 2.95447\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 5.8649 - acc: 0.5086 - val_loss: 3.3945 - val_acc: 0.5390\n",
      "Epoch 6/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 3.7057 - acc: 0.5092\n",
      "Epoch 6: val_loss improved from 2.95447 to 2.21444, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 15ms/step - loss: 3.7273 - acc: 0.5086 - val_loss: 2.2144 - val_acc: 0.4705\n",
      "Epoch 7/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 2.2903 - acc: 0.5000\n",
      "Epoch 7: val_loss improved from 2.21444 to 1.44343, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 12ms/step - loss: 2.2829 - acc: 0.4975 - val_loss: 1.4434 - val_acc: 0.5190\n",
      "Epoch 8/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 1.4664 - acc: 0.5210\n",
      "Epoch 8: val_loss improved from 1.44343 to 1.28689, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 1.4682 - acc: 0.5218 - val_loss: 1.2869 - val_acc: 0.5116\n",
      "Epoch 9/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 1.2422 - acc: 0.5676\n",
      "Epoch 9: val_loss improved from 1.28689 to 1.23223, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 12ms/step - loss: 1.2419 - acc: 0.5661 - val_loss: 1.2322 - val_acc: 0.4568\n",
      "Epoch 10/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.2984 - acc: 0.5631\n",
      "Epoch 10: val_loss improved from 1.23223 to 1.02553, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 1.2974 - acc: 0.5700 - val_loss: 1.0255 - val_acc: 0.6435\n",
      "Epoch 11/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 1.2302 - acc: 0.5959\n",
      "Epoch 11: val_loss improved from 1.02553 to 0.97322, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 12ms/step - loss: 1.2310 - acc: 0.5980 - val_loss: 0.9732 - val_acc: 0.6044\n",
      "Epoch 12/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.1146 - acc: 0.6177\n",
      "Epoch 12: val_loss did not improve from 0.97322\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.1226 - acc: 0.6180 - val_loss: 0.9811 - val_acc: 0.6076\n",
      "Epoch 13/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.1324 - acc: 0.6032\n",
      "Epoch 13: val_loss did not improve from 0.97322\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.1222 - acc: 0.6109 - val_loss: 1.0536 - val_acc: 0.6572\n",
      "Epoch 14/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 1.0714 - acc: 0.6479\n",
      "Epoch 14: val_loss did not improve from 0.97322\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 1.0995 - acc: 0.6463 - val_loss: 1.1143 - val_acc: 0.5654\n",
      "Epoch 15/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.2667 - acc: 0.6242\n",
      "Epoch 15: val_loss did not improve from 0.97322\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.2676 - acc: 0.6260 - val_loss: 1.2554 - val_acc: 0.6508\n",
      "Epoch 16/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.0815 - acc: 0.6621\n",
      "Epoch 16: val_loss did not improve from 0.97322\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 1.0919 - acc: 0.6568 - val_loss: 1.3299 - val_acc: 0.5844\n",
      "Epoch 17/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.0784 - acc: 0.6730\n",
      "Epoch 17: val_loss improved from 0.97322 to 0.92692, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 12ms/step - loss: 1.0623 - acc: 0.6758 - val_loss: 0.9269 - val_acc: 0.6804\n",
      "Epoch 18/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 1.0553 - acc: 0.6627\n",
      "Epoch 18: val_loss did not improve from 0.92692\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.0584 - acc: 0.6626 - val_loss: 1.1298 - val_acc: 0.5759\n",
      "Epoch 19/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.1059 - acc: 0.6585\n",
      "Epoch 19: val_loss did not improve from 0.92692\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.1183 - acc: 0.6494 - val_loss: 1.0145 - val_acc: 0.5696\n",
      "Epoch 20/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.0269 - acc: 0.6749\n",
      "Epoch 20: val_loss did not improve from 0.92692\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.0337 - acc: 0.6742 - val_loss: 1.0924 - val_acc: 0.7025\n",
      "Epoch 21/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 1.0665 - acc: 0.6707\n",
      "Epoch 21: val_loss did not improve from 0.92692\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.0634 - acc: 0.6729 - val_loss: 1.0939 - val_acc: 0.6772\n",
      "Epoch 22/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 1.0214 - acc: 0.6794\n",
      "Epoch 22: val_loss did not improve from 0.92692\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.0280 - acc: 0.6755 - val_loss: 0.9687 - val_acc: 0.6793\n",
      "Epoch 23/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 1.0048 - acc: 0.6861\n",
      "Epoch 23: val_loss improved from 0.92692 to 0.87484, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 12ms/step - loss: 0.9974 - acc: 0.6864 - val_loss: 0.8748 - val_acc: 0.7257\n",
      "Epoch 24/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 0.9858 - acc: 0.7014\n",
      "Epoch 24: val_loss did not improve from 0.87484\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 0.9802 - acc: 0.7011 - val_loss: 0.9774 - val_acc: 0.6329\n",
      "Epoch 25/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.0062 - acc: 0.6917\n",
      "Epoch 25: val_loss did not improve from 0.87484\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.0083 - acc: 0.6893 - val_loss: 1.2583 - val_acc: 0.5253\n",
      "Epoch 26/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 1.1180 - acc: 0.6805\n",
      "Epoch 26: val_loss did not improve from 0.87484\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.1239 - acc: 0.6806 - val_loss: 0.9927 - val_acc: 0.6424\n",
      "Epoch 27/100\n",
      "26/30 [=========================>....] - ETA: 0s - loss: 1.0382 - acc: 0.6896\n",
      "Epoch 27: val_loss did not improve from 0.87484\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 1.0737 - acc: 0.6908 - val_loss: 0.9043 - val_acc: 0.6899\n",
      "Epoch 28/100\n",
      "26/30 [=========================>....] - ETA: 0s - loss: 0.9676 - acc: 0.7028\n",
      "Epoch 28: val_loss improved from 0.87484 to 0.87325, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 12ms/step - loss: 0.9708 - acc: 0.6948 - val_loss: 0.8733 - val_acc: 0.7205\n",
      "Epoch 29/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.9919 - acc: 0.7085\n",
      "Epoch 29: val_loss did not improve from 0.87325\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 0.9931 - acc: 0.7085 - val_loss: 0.8847 - val_acc: 0.6603\n",
      "Epoch 30/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.9580 - acc: 0.6967\n",
      "Epoch 30: val_loss did not improve from 0.87325\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 0.9612 - acc: 0.6988 - val_loss: 0.9405 - val_acc: 0.7173\n",
      "Epoch 31/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.9210 - acc: 0.7117\n",
      "Epoch 31: val_loss did not improve from 0.87325\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "30/30 [==============================] - 0s 10ms/step - loss: 0.9210 - acc: 0.7117 - val_loss: 0.9005 - val_acc: 0.6572\n",
      "Epoch 32/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.9355 - acc: 0.7065\n",
      "Epoch 32: val_loss did not improve from 0.87325\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 0.9652 - acc: 0.6993 - val_loss: 1.4843 - val_acc: 0.5717\n",
      "Epoch 33/100\n",
      "24/30 [=======================>......] - ETA: 0s - loss: 1.3448 - acc: 0.6533\n",
      "Epoch 33: val_loss did not improve from 0.87325\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.3373 - acc: 0.6463 - val_loss: 1.8147 - val_acc: 0.6709\n",
      "Epoch 34/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.1783 - acc: 0.6861\n",
      "Epoch 34: val_loss did not improve from 0.87325\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.1711 - acc: 0.6811 - val_loss: 0.9740 - val_acc: 0.6118\n",
      "Epoch 35/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 1.1250 - acc: 0.6762\n",
      "Epoch 35: val_loss did not improve from 0.87325\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.1221 - acc: 0.6750 - val_loss: 0.9604 - val_acc: 0.6867\n",
      "Epoch 36/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 0.9162 - acc: 0.7138\n",
      "Epoch 36: val_loss did not improve from 0.87325\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 0.9226 - acc: 0.7114 - val_loss: 1.1350 - val_acc: 0.6857\n",
      "Epoch 37/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.0451 - acc: 0.6825\n",
      "Epoch 37: val_loss did not improve from 0.87325\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.0534 - acc: 0.6840 - val_loss: 0.9046 - val_acc: 0.7025\n",
      "Epoch 38/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 1.0454 - acc: 0.6814\n",
      "Epoch 38: val_loss improved from 0.87325 to 0.85698, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 1.0732 - acc: 0.6850 - val_loss: 0.8570 - val_acc: 0.7278\n",
      "Epoch 39/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.9612 - acc: 0.6970\n",
      "Epoch 39: val_loss did not improve from 0.85698\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 0.9610 - acc: 0.6972 - val_loss: 0.8928 - val_acc: 0.6677\n",
      "Epoch 40/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.9977 - acc: 0.7012\n",
      "Epoch 40: val_loss did not improve from 0.85698\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 0.9880 - acc: 0.7046 - val_loss: 1.0375 - val_acc: 0.6814\n",
      "Epoch 41/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.0623 - acc: 0.6925\n",
      "Epoch 41: val_loss did not improve from 0.85698\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.0435 - acc: 0.6959 - val_loss: 0.9065 - val_acc: 0.7226\n",
      "Epoch 42/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.9994 - acc: 0.7091\n",
      "Epoch 42: val_loss did not improve from 0.85698\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.0023 - acc: 0.7088 - val_loss: 0.9760 - val_acc: 0.6846\n",
      "Epoch 43/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.9462 - acc: 0.7062\n",
      "Epoch 43: val_loss improved from 0.85698 to 0.81848, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 1s 30ms/step - loss: 0.9545 - acc: 0.7067 - val_loss: 0.8185 - val_acc: 0.7300\n",
      "Epoch 44/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.8800 - acc: 0.7201\n",
      "Epoch 44: val_loss improved from 0.81848 to 0.79675, saving model to dips_weights.h5\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 0.8806 - acc: 0.7201 - val_loss: 0.7967 - val_acc: 0.7384\n",
      "Epoch 45/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.8370 - acc: 0.7344\n",
      "Epoch 45: val_loss did not improve from 0.79675\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 0.8320 - acc: 0.7365 - val_loss: 0.8029 - val_acc: 0.7321\n",
      "Epoch 46/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.0254 - acc: 0.6869\n",
      "Epoch 46: val_loss did not improve from 0.79675\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.0612 - acc: 0.6866 - val_loss: 1.6523 - val_acc: 0.6688\n",
      "Epoch 47/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.3785 - acc: 0.6465\n",
      "Epoch 47: val_loss did not improve from 0.79675\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.3572 - acc: 0.6471 - val_loss: 1.0101 - val_acc: 0.7068\n",
      "Epoch 48/100\n",
      "23/30 [======================>.......] - ETA: 0s - loss: 1.0255 - acc: 0.6957\n",
      "Epoch 48: val_loss did not improve from 0.79675\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 0.9782 - acc: 0.7069 - val_loss: 0.8719 - val_acc: 0.6846\n",
      "Epoch 49/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.9442 - acc: 0.6975\n",
      "Epoch 49: val_loss did not improve from 0.79675\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 0.9472 - acc: 0.6966 - val_loss: 0.8352 - val_acc: 0.7236\n",
      "Epoch 50/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.9418 - acc: 0.7055\n",
      "Epoch 50: val_loss did not improve from 0.79675\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 0.9475 - acc: 0.7040 - val_loss: 0.9475 - val_acc: 0.7099\n",
      "Epoch 51/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.8985 - acc: 0.7185\n",
      "Epoch 51: val_loss did not improve from 0.79675\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 0.8967 - acc: 0.7188 - val_loss: 0.9651 - val_acc: 0.6751\n",
      "Epoch 52/100\n",
      "26/30 [=========================>....] - ETA: 0s - loss: 1.0624 - acc: 0.6830\n",
      "Epoch 52: val_loss did not improve from 0.79675\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 1.0463 - acc: 0.6861 - val_loss: 0.9257 - val_acc: 0.7131\n",
      "Epoch 53/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.9398 - acc: 0.7015\n",
      "Epoch 53: val_loss did not improve from 0.79675\n",
      "30/30 [==============================] - 0s 11ms/step - loss: 0.9390 - acc: 0.7027 - val_loss: 0.9172 - val_acc: 0.6962\n",
      "Epoch 54/100\n",
      "26/30 [=========================>....] - ETA: 0s - loss: 1.0425 - acc: 0.6884\n",
      "Epoch 54: val_loss did not improve from 0.79675\n",
      "30/30 [==============================] - 0s 10ms/step - loss: 1.0480 - acc: 0.6861 - val_loss: 1.3513 - val_acc: 0.6909\n",
      "Epoch 54: early stopping\n"
     ]
    }
   ],
   "source": [
    "opt = keras.optimizers.Adam(learning_rate=0.0001)\n",
    "\n",
    "dips_mean.compile(loss='categorical_crossentropy', optimizer=opt,\n",
    "                 metrics=['acc'])  \n",
    "\n",
    "dips_hist_mean = dips_mean.fit(X_train, y_train_cat, epochs=nEpochs, \n",
    "                     batch_size=128,validation_split=0.2,\n",
    "                     callbacks=[earlyStop, dips_mChkPt])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "55dd0aa6-8d6a-4c8d-991a-2a5bd5f1deb4",
   "metadata": {},
   "source": [
    "**Which of these 3 models is doing the best?** "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "46f7f3fd-55aa-4c33-a6f6-ce0b08360b54",
   "metadata": {},
   "source": [
    "Initial check: Compare the loss curves."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "c3666096-f2c5-4443-8c81-d15a63203781",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for h,c,l in zip([dips_hist, dips_hist_max, dips_hist_mean],\n",
    "                 ['hotpink','C2','C1'],['Sum','Max','Mean']):\n",
    "\n",
    "    epochs = np.arange(1,len(h.history['loss'])+1)\n",
    "    \n",
    "    plt.plot(epochs,h.history['loss'],color=c,label=f'{l} : train')\n",
    "    plt.plot(epochs,h.history['val_loss'],color=c,label=f'{l} : val',ls='--')\n",
    "    \n",
    "plt.xlabel('epochs',fontsize=14)\n",
    "plt.ylabel('cross-entropy loss',fontsize=14)\n",
    "plt.legend()\n",
    "plt.title('DIPS: Different permutation operations')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a6278d68-ea98-43ec-9159-3aad7f364103",
   "metadata": {},
   "source": [
    "Next: compare the roc curves on the test set"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "f4389123-407a-484c-9962-792883f2f5f2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "79/79 [==============================] - 0s 3ms/step\n",
      "Db discriminant range: -64.807434 85.67963\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_7268/2548379725.py:23: RuntimeWarning: divide by zero encountered in true_divide\n",
      "  disc = np.log(np.divide(predictions[:,2], fc*predictions[:,1] + (1 - fc) * predictions[:,0]))\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "leff_max, ceff_max, beff_max = sigBkgEff(dips_max, X_test, y_test, \n",
    "                                         title='Max pooling operation')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "69c4c774-7345-4380-bf5a-6934a5f6af49",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "79/79 [==============================] - 0s 3ms/step\n",
      "Db discriminant range: -54.238068 31.93204\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "leff_mean, ceff_mean, beff_mean = sigBkgEff(dips_mean, X_test, y_test, \n",
    "                                            title='Mean (avg) pooling operation')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "2ae93fdb-0f9a-4a76-953a-8052b7dde29e",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_7268/4292196905.py:7: RuntimeWarning: divide by zero encountered in true_divide\n",
      "  plt.plot(bi, 1 / li, color=c, label=f'{l}: l-rej')\n",
      "/tmp/ipykernel_7268/4292196905.py:7: RuntimeWarning: divide by zero encountered in true_divide\n",
      "  plt.plot(bi, 1 / li, color=c, label=f'{l}: l-rej')\n",
      "/tmp/ipykernel_7268/4292196905.py:10: RuntimeWarning: divide by zero encountered in true_divide\n",
      "  plt.plot(bi, 1 / ci, color=c, linestyle='--', label=f'{l}: c-rej')\n",
      "/tmp/ipykernel_7268/4292196905.py:18: UserWarning: Attempted to set non-positive bottom ylim on a log-scaled axis.\n",
      "Invalid limit will be ignored.\n",
      "  plt.ylim(0,int(2e3))\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure()\n",
    "\n",
    "for bi,li,ci,c, l in zip([beff,beff_max,beff_mean],[leff,leff_max,leff_mean],\n",
    "                         [ceff,ceff_max,ceff_mean],['hotpink','C2','C1'],['Sum','Max','Mean']):\n",
    "\n",
    "    # l-rej\n",
    "    plt.plot(bi, 1 / li, color=c, label=f'{l}: l-rej')\n",
    "\n",
    "    # c-rej\n",
    "    plt.plot(bi, 1 / ci, color=c, linestyle='--', label=f'{l}: c-rej')\n",
    "plt.xlabel('b efficiency')\n",
    "plt.ylabel('Background rejection')\n",
    "\n",
    "plt.legend()\n",
    "#plt.title(title)\n",
    "plt.yscale(\"log\")\n",
    "plt.xlim(0.6,1)\n",
    "plt.ylim(0,int(2e3))\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "78142c2f-5a4b-404a-875f-657ba29b4054",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "df48150b-9be9-4f40-80eb-745c709114cc",
   "metadata": {},
   "source": [
    "**Q1:** Which is the least performant model?\n",
    "\n",
    "-> The Max pooling operation \n",
    "\n",
    "**Q2** Why might this model be doing not as good?\n",
    "\n",
    "-> By just taking the max over the features, we're loosing information from the features for many of the jets."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0b34554b-22a2-4df2-bc9b-91eecc420418",
   "metadata": {},
   "outputs": [],
   "source": [
    " "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7209c8b7-6a22-4464-bd62-c37efe4fa7d1",
   "metadata": {},
   "source": [
    "**Step 4:** Test out batch norm\n",
    "<a name=\"bn\"></a>\n",
    "\n",
    "[BatchNormalization](https://arxiv.org/abs/1502.03167) is a regularization tecnique for deep models. The key idea is that it's easier for models to train if the input are bell shaped and centered around 0 to take advantage of the non-linearity. This is why we normalize the inputs in the pre-processing, and the Batch-Normalization is a technique which normalizes the outputs from a network layer so that the next layer has nicely scaled inputs for the training. "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c3c7665-4bc0-4bf2-b1e7-155cedcb366c",
   "metadata": {},
   "source": [
    "**Add batch norm to the model**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "id": "d19ba77b-0d0f-4cd9-8e6e-12a037fca052",
   "metadata": {},
   "outputs": [],
   "source": [
    "trk_inputs = Input(shape=(maxNumTrks,nFeatures),name='Input')\n",
    "masked_inputs = Masking(mask_value=0)(trk_inputs)\n",
    "tdd = masked_inputs\n",
    "\n",
    "for i, phi_nodes in enumerate(ppm_sizes_int):\n",
    "\n",
    "    tdd = TimeDistributed(Dense(phi_nodes,activation='linear'),name=f\"Phi{i}_Dense\")(tdd)\n",
    "\n",
    "    # Example - how to add batch norm to the Phi network\n",
    "    tdd = TimeDistributed(BatchNormalization(),name=f\"Phi{i}_BatchNormalization\")(tdd)\n",
    "    \n",
    "    tdd = TimeDistributed(ReLU(),name=f\"Phi{i}_ReLU\")(tdd)\n",
    "\n",
    "# This is where the magic happens... sum up the track features!\n",
    "F = Sum(name=\"Sum\")(tdd)\n",
    "\n",
    "for j, F_nodes in enumerate(dense_sizes_int):\n",
    "\n",
    "    F = Dense(F_nodes, activation='linear', name=f\"F{j}_Dense\")(F)\n",
    "    \n",
    "    '''\n",
    "    TO DO: Also add batch norm to the F dense network\n",
    "    '''\n",
    "    F = BatchNormalization(name=f\"F{j}_BatchNormalization\")(F)\n",
    "    #raise NotImplementedError\n",
    "    \n",
    "    F = ReLU(name=f\"F{j}_ReLU\")(F)\n",
    "\n",
    "output = Dense(nClasses, activation='softmax',name=\"Jet_class\")(F)\n",
    "dips_bn = Model(inputs=trk_inputs, outputs=output)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "62433c73-7969-4554-ae44-2f0e9473c6db",
   "metadata": {},
   "source": [
    "**Train the model**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "id": "53640911-fd89-4352-9332-af265804e9dd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 1.1964 - acc: 0.3513\n",
      "Epoch 1: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 4s 37ms/step - loss: 1.1919 - acc: 0.3545 - val_loss: 1.2078 - val_acc: 0.1825\n",
      "Epoch 2/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 1.0574 - acc: 0.4461\n",
      "Epoch 2: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 1.0574 - acc: 0.4461 - val_loss: 1.1417 - val_acc: 0.3302\n",
      "Epoch 3/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.9772 - acc: 0.5431\n",
      "Epoch 3: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 0.9772 - acc: 0.5431 - val_loss: 1.1103 - val_acc: 0.4030\n",
      "Epoch 4/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.9403 - acc: 0.5811\n",
      "Epoch 4: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 0.9403 - acc: 0.5811 - val_loss: 1.0826 - val_acc: 0.5348\n",
      "Epoch 5/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 0.8990 - acc: 0.6218\n",
      "Epoch 5: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.8942 - acc: 0.6260 - val_loss: 1.0558 - val_acc: 0.5612\n",
      "Epoch 6/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.8653 - acc: 0.6626\n",
      "Epoch 6: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.8653 - acc: 0.6626 - val_loss: 1.0052 - val_acc: 0.5781\n",
      "Epoch 7/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.8427 - acc: 0.6730\n",
      "Epoch 7: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 21ms/step - loss: 0.8422 - acc: 0.6729 - val_loss: 0.9705 - val_acc: 0.5823\n",
      "Epoch 8/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.8133 - acc: 0.6943\n",
      "Epoch 8: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.8133 - acc: 0.6943 - val_loss: 0.9457 - val_acc: 0.5907\n",
      "Epoch 9/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.8004 - acc: 0.6929\n",
      "Epoch 9: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.8017 - acc: 0.6908 - val_loss: 0.9241 - val_acc: 0.6013\n",
      "Epoch 10/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.7747 - acc: 0.7132\n",
      "Epoch 10: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.7785 - acc: 0.7114 - val_loss: 0.9118 - val_acc: 0.6086\n",
      "Epoch 11/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.7662 - acc: 0.7119\n",
      "Epoch 11: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.7662 - acc: 0.7119 - val_loss: 0.8850 - val_acc: 0.6181\n",
      "Epoch 12/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.7545 - acc: 0.7147\n",
      "Epoch 12: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.7550 - acc: 0.7146 - val_loss: 0.8587 - val_acc: 0.6329\n",
      "Epoch 13/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.7419 - acc: 0.7241\n",
      "Epoch 13: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.7419 - acc: 0.7241 - val_loss: 0.8534 - val_acc: 0.6435\n",
      "Epoch 14/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.7308 - acc: 0.7247\n",
      "Epoch 14: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.7306 - acc: 0.7249 - val_loss: 0.8557 - val_acc: 0.6445\n",
      "Epoch 15/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.7281 - acc: 0.7251\n",
      "Epoch 15: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.7281 - acc: 0.7251 - val_loss: 0.8447 - val_acc: 0.6593\n",
      "Epoch 16/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.7109 - acc: 0.7344\n",
      "Epoch 16: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.7109 - acc: 0.7344 - val_loss: 0.8367 - val_acc: 0.6614\n",
      "Epoch 17/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.7112 - acc: 0.7321\n",
      "Epoch 17: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.7107 - acc: 0.7328 - val_loss: 0.7768 - val_acc: 0.6920\n",
      "Epoch 18/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6939 - acc: 0.7419\n",
      "Epoch 18: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6957 - acc: 0.7402 - val_loss: 0.7781 - val_acc: 0.6846\n",
      "Epoch 19/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6865 - acc: 0.7443\n",
      "Epoch 19: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6889 - acc: 0.7431 - val_loss: 0.7733 - val_acc: 0.6835\n",
      "Epoch 20/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.6847 - acc: 0.7447\n",
      "Epoch 20: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6847 - acc: 0.7447 - val_loss: 0.7784 - val_acc: 0.6878\n",
      "Epoch 21/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6785 - acc: 0.7492\n",
      "Epoch 21: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6768 - acc: 0.7502 - val_loss: 0.7385 - val_acc: 0.7152\n",
      "Epoch 22/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 0.6663 - acc: 0.7561\n",
      "Epoch 22: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6713 - acc: 0.7526 - val_loss: 0.7351 - val_acc: 0.7152\n",
      "Epoch 23/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.6708 - acc: 0.7494\n",
      "Epoch 23: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6708 - acc: 0.7494 - val_loss: 0.7328 - val_acc: 0.7068\n",
      "Epoch 24/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6622 - acc: 0.7524\n",
      "Epoch 24: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6611 - acc: 0.7534 - val_loss: 0.7177 - val_acc: 0.7342\n",
      "Epoch 25/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.6598 - acc: 0.7564\n",
      "Epoch 25: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6600 - acc: 0.7568 - val_loss: 0.7154 - val_acc: 0.7278\n",
      "Epoch 26/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.6545 - acc: 0.7536\n",
      "Epoch 26: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6533 - acc: 0.7547 - val_loss: 0.7055 - val_acc: 0.7426\n",
      "Epoch 27/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6444 - acc: 0.7637\n",
      "Epoch 27: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6470 - acc: 0.7621 - val_loss: 0.6936 - val_acc: 0.7363\n",
      "Epoch 28/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.6456 - acc: 0.7586\n",
      "Epoch 28: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6436 - acc: 0.7597 - val_loss: 0.6942 - val_acc: 0.7447\n",
      "Epoch 29/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6433 - acc: 0.7589\n",
      "Epoch 29: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6446 - acc: 0.7589 - val_loss: 0.6779 - val_acc: 0.7479\n",
      "Epoch 30/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.6424 - acc: 0.7603\n",
      "Epoch 30: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6413 - acc: 0.7618 - val_loss: 0.6879 - val_acc: 0.7416\n",
      "Epoch 31/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6385 - acc: 0.7643\n",
      "Epoch 31: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6439 - acc: 0.7623 - val_loss: 0.6760 - val_acc: 0.7500\n",
      "Epoch 32/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.6366 - acc: 0.7695\n",
      "Epoch 32: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6366 - acc: 0.7695 - val_loss: 0.6790 - val_acc: 0.7416\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 33/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.6335 - acc: 0.7709\n",
      "Epoch 33: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6333 - acc: 0.7702 - val_loss: 0.6832 - val_acc: 0.7489\n",
      "Epoch 34/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6297 - acc: 0.7707\n",
      "Epoch 34: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6304 - acc: 0.7708 - val_loss: 0.6753 - val_acc: 0.7447\n",
      "Epoch 35/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6339 - acc: 0.7667\n",
      "Epoch 35: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6349 - acc: 0.7660 - val_loss: 0.6642 - val_acc: 0.7595\n",
      "Epoch 36/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.6337 - acc: 0.7704\n",
      "Epoch 36: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6354 - acc: 0.7681 - val_loss: 0.6688 - val_acc: 0.7521\n",
      "Epoch 37/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6261 - acc: 0.7664\n",
      "Epoch 37: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6288 - acc: 0.7650 - val_loss: 0.6733 - val_acc: 0.7489\n",
      "Epoch 38/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.6315 - acc: 0.7600\n",
      "Epoch 38: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6315 - acc: 0.7600 - val_loss: 0.6671 - val_acc: 0.7563\n",
      "Epoch 39/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6303 - acc: 0.7654\n",
      "Epoch 39: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6291 - acc: 0.7663 - val_loss: 0.6586 - val_acc: 0.7553\n",
      "Epoch 40/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6332 - acc: 0.7716\n",
      "Epoch 40: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6352 - acc: 0.7700 - val_loss: 0.6952 - val_acc: 0.7447\n",
      "Epoch 41/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.6279 - acc: 0.7710\n",
      "Epoch 41: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6279 - acc: 0.7710 - val_loss: 0.6569 - val_acc: 0.7595\n",
      "Epoch 42/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 0.6366 - acc: 0.7665\n",
      "Epoch 42: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6294 - acc: 0.7684 - val_loss: 0.6523 - val_acc: 0.7627\n",
      "Epoch 43/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6295 - acc: 0.7718\n",
      "Epoch 43: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6327 - acc: 0.7702 - val_loss: 0.6570 - val_acc: 0.7563\n",
      "Epoch 44/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.6230 - acc: 0.7720\n",
      "Epoch 44: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6258 - acc: 0.7705 - val_loss: 0.6544 - val_acc: 0.7605\n",
      "Epoch 45/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 0.6242 - acc: 0.7700\n",
      "Epoch 45: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6246 - acc: 0.7700 - val_loss: 0.6514 - val_acc: 0.7563\n",
      "Epoch 46/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.6263 - acc: 0.7700\n",
      "Epoch 46: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6263 - acc: 0.7700 - val_loss: 0.6498 - val_acc: 0.7616\n",
      "Epoch 47/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6264 - acc: 0.7672\n",
      "Epoch 47: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 22ms/step - loss: 0.6271 - acc: 0.7668 - val_loss: 0.6602 - val_acc: 0.7532\n",
      "Epoch 48/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6154 - acc: 0.7775\n",
      "Epoch 48: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6192 - acc: 0.7763 - val_loss: 0.6567 - val_acc: 0.7595\n",
      "Epoch 49/100\n",
      "30/30 [==============================] - ETA: 0s - loss: 0.6279 - acc: 0.7652\n",
      "Epoch 49: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6279 - acc: 0.7652 - val_loss: 0.6638 - val_acc: 0.7595\n",
      "Epoch 50/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.6245 - acc: 0.7718\n",
      "Epoch 50: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 18ms/step - loss: 0.6230 - acc: 0.7721 - val_loss: 0.6534 - val_acc: 0.7542\n",
      "Epoch 51/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.6272 - acc: 0.7720\n",
      "Epoch 51: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6243 - acc: 0.7729 - val_loss: 0.6541 - val_acc: 0.7532\n",
      "Epoch 52/100\n",
      "27/30 [==========================>...] - ETA: 0s - loss: 0.6236 - acc: 0.7694\n",
      "Epoch 52: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6229 - acc: 0.7697 - val_loss: 0.6528 - val_acc: 0.7595\n",
      "Epoch 53/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6252 - acc: 0.7659\n",
      "Epoch 53: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6224 - acc: 0.7684 - val_loss: 0.6602 - val_acc: 0.7532\n",
      "Epoch 54/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.6098 - acc: 0.7799\n",
      "Epoch 54: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6147 - acc: 0.7768 - val_loss: 0.6585 - val_acc: 0.7521\n",
      "Epoch 55/100\n",
      "29/30 [============================>.] - ETA: 0s - loss: 0.6211 - acc: 0.7694\n",
      "Epoch 55: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 19ms/step - loss: 0.6215 - acc: 0.7695 - val_loss: 0.6528 - val_acc: 0.7616\n",
      "Epoch 56/100\n",
      "28/30 [===========================>..] - ETA: 0s - loss: 0.6249 - acc: 0.7701\n",
      "Epoch 56: val_loss did not improve from 0.64669\n",
      "30/30 [==============================] - 1s 20ms/step - loss: 0.6233 - acc: 0.7713 - val_loss: 0.6526 - val_acc: 0.7532\n",
      "Epoch 56: early stopping\n"
     ]
    }
   ],
   "source": [
    "opt = keras.optimizers.Adam(learning_rate=0.0001)\n",
    "\n",
    "\n",
    "dips_bn.compile(loss='categorical_crossentropy', optimizer=opt,\n",
    "                 metrics=['acc'])  \n",
    "\n",
    "dips_hist_bn = dips_bn.fit(X_train, y_train_cat, epochs=nEpochs, \n",
    "                         batch_size=128,validation_split=0.2,\n",
    "                         callbacks=[earlyStop, dips_mChkPt])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "845dc966-82e7-4568-9b8f-69a1a564152c",
   "metadata": {},
   "source": [
    "**Compare the training curves**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "id": "3ecbdc60-f2b7-48d5-8319-f413696c83cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for h,c,l in zip([dips_hist, dips_hist_bn],\n",
    "                 ['hotpink','navy'],['default','BatchNorm']):\n",
    "    \n",
    "    epochs = np.arange(1,len(h.history['loss'])+1)\n",
    "    \n",
    "    plt.plot(epochs,h.history['loss'],color=c,label=f'{l} : train')\n",
    "    plt.plot(epochs,h.history['val_loss'],color=c,label=f'{l} : val',ls='--')\n",
    "    \n",
    "plt.xlabel('epochs',fontsize=14)\n",
    "plt.ylabel('cross-entropy loss',fontsize=14)\n",
    "plt.legend()\n",
    "plt.title('DIPS: Impact of batch norm')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ef781bb5-c5cf-48b9-847f-5cb335b9f40e",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "id": "fe3a738e-b514-47cb-b950-34bc678b88cb",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "79/79 [==============================] - 1s 3ms/step\n",
      "Db discriminant range: -2.7810357 9.476455\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "leff_bn, ceff_bn, beff_bn = sigBkgEff(dips_bn, X_test, y_test,  \n",
    "                                      title='Adding batch norm')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "id": "b334f971-eb5e-49a6-b606-65fa2718acd3",
   "metadata": {},
   "outputs": [],
   "source": [
    "from scipy.interpolate import pchip"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "id": "83023098-7fd8-4fd9-b07b-449a804052ca",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/tmp/ipykernel_7268/2044709358.py:8: RuntimeWarning: divide by zero encountered in true_divide\n",
      "  ax1.plot(bi, 1 / li, color=c, label=f'{l}: l-rej')\n",
      "/tmp/ipykernel_7268/2044709358.py:10: RuntimeWarning: divide by zero encountered in true_divide\n",
      "  ax1.plot(bi, 1 / ci, color=c, linestyle='--', label=f'{l}: c-rej')\n",
      "/tmp/ipykernel_7268/2044709358.py:38: UserWarning: Attempted to set non-positive bottom ylim on a log-scaled axis.\n",
      "Invalid limit will be ignored.\n",
      "  ax1.set_ylim(0,int(2e3))\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x432 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, (ax1,ax2) = plt.subplots(2,1,figsize=(6,6),sharex=True,\n",
    "                              gridspec_kw={'height_ratios':[2,1]})\n",
    "\n",
    "for bi,li,ci,c, l in zip([beff,beff_bn],[leff,leff_bn],[ceff,ceff_bn],\n",
    "                         ['hotpink','navy'],['default','BatchNorm']):\n",
    "\n",
    "    # l-rej\n",
    "    ax1.plot(bi, 1 / li, color=c, label=f'{l}: l-rej')\n",
    "    # c-rej\n",
    "    ax1.plot(bi, 1 / ci, color=c, linestyle='--', label=f'{l}: c-rej')\n",
    "    \n",
    "    \n",
    "# Also the ratio panel\n",
    "xx = np.linspace(0.6,1,101)\n",
    "\n",
    "dx = np.concatenate((np.ones(1),np.diff(beff)))\n",
    "dx_bn = np.concatenate((np.ones(1),np.diff(beff_bn)))\n",
    "\n",
    "# l-rej\n",
    "for bkg,bkg_bn, ls in zip([leff,ceff],[leff_bn,ceff_bn],['-','--']):\n",
    "    \n",
    "    m_num = (bkg_bn!=0) & (dx_bn>0)\n",
    "    m_den = (bkg!=0) & (dx>0)\n",
    "    \n",
    "    f_num = pchip(beff_bn[m_num], 1/bkg_bn[m_num]) \n",
    "    f_den = pchip(beff[m_den],    1/bkg[m_den]) \n",
    "\n",
    "    ax2.plot(xx, f_num(xx) / f_den(xx), ls=ls, color=c)\n",
    "    \n",
    "ax2.set_xlabel('b efficiency')\n",
    "ax1.set_ylabel('Background rejection')\n",
    "ax1.set_ylabel('Background rejection')\n",
    "\n",
    "ax1.legend()\n",
    "ax1.set_title(title)\n",
    "ax1.set_yscale(\"log\")\n",
    "ax1.set_xlim(0.6,1)\n",
    "ax1.set_ylim(0,int(2e3))\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2efb704c-6917-47ad-b4c4-0fb7ba4e8231",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "fa8f900f-bfe0-49e2-a804-e06d5c09f0e3",
   "metadata": {},
   "source": [
    "**Look at the activation functions**\n",
    "\n",
    "Let's start off by looking at the first layer of the $\\Phi$ network."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "id": "408fd1f5-b97f-4e89-bf49-1b4c0909b7e3",
   "metadata": {},
   "outputs": [],
   "source": [
    "phi_i = 0\n",
    "\n",
    "ni = 10000"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "id": "509bc17e-8e4c-47ab-ba91-0949b5cd9c9f",
   "metadata": {},
   "outputs": [],
   "source": [
    "func = K.function([dips.get_layer('Input').input], dips.get_layer(f'Phi{phi_i}_Dense').output)\n",
    "act = func([X_test[:ni]])  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "id": "2dca452b-35d4-4e8d-9f0c-a14daee394e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "To do - fill in to get the activations from the batch norm trained network\n",
    "'''\n",
    "\n",
    "func = K.function([dips_bn.get_layer('Input').input], dips_bn.get_layer(f'Phi{phi_i}_Dense').output)\n",
    "bnIn = func([X_test[:ni]])  \n",
    "\n",
    "func = K.function([dips_bn.get_layer('Input').input], dips_bn.get_layer(f'Phi{phi_i}_BatchNormalization').output)\n",
    "bnOut = func([X_test[:ni]])  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "60fbfc98-aec8-4c4a-95a9-bbc2f6425488",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "1a60567d-d41b-4c57-a760-5d85a3972e66",
   "metadata": {},
   "source": [
    "To avoid evaluating the activations that _don't_ have any tracks, need to define the track mask"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "id": "0653ec2d-337f-4116-b8a1-9e2212b8b992",
   "metadata": {},
   "outputs": [],
   "source": [
    "tmask = ~np.all(X_test[:ni]==0, axis=-1) # False for a padded track"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "id": "7087705a-e819-4ded-8dae-65241d4204f3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2507, 40)"
      ]
     },
     "execution_count": 109,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmask.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "129a3a6e-8494-4726-8f75-bcc3bd6407ca",
   "metadata": {},
   "source": [
    "This track mask has shape (nJets, maxNumTrks)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "id": "c3f9846f-ca33-4931-9e4d-8444bff102b4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([[ True,  True,  True,  True,  True,  True,  True,  True,  True,\n",
       "         True,  True,  True,  True,  True,  True,  True,  True,  True,\n",
       "         True,  True,  True,  True,  True,  True, False, False, False,\n",
       "        False, False, False, False, False, False, False, False, False,\n",
       "        False, False, False, False],\n",
       "       [ True,  True,  True,  True,  True,  True,  True,  True,  True,\n",
       "         True,  True,  True,  True,  True,  True,  True,  True,  True,\n",
       "         True,  True,  True,  True,  True,  True,  True,  True,  True,\n",
       "         True,  True,  True,  True, False, False, False, False, False,\n",
       "        False, False, False, False]])"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmask[:2]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e6c5b653-9620-43b1-a3cc-57ff6d1615ab",
   "metadata": {},
   "source": [
    "And looking at the first two jets we can see:\n",
    "- The first jet has 6 tracks\n",
    "- The 2nd jet has 5 tracks"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "id": "37dea649-5f43-4ad9-bb6d-488651b9c106",
   "metadata": {},
   "outputs": [],
   "source": [
    "from matplotlib import transforms\n",
    "\n",
    "def rainbow_text(x,y,ls,lc,yoffset=.75,ax=None,**kw):\n",
    "    '''\n",
    "    Take a list of strings ``ls`` and colors ``lc`` and place them next to each\n",
    "    other, with text ls[i] being shown in color lc[i].\n",
    "\n",
    "    This example shows how to do both vertical and horizontal text, and will\n",
    "    pass all keyword arguments to plt.text, so you can set the font size,\n",
    "    family, etc.\n",
    "    '''\n",
    "\n",
    "    if ax is None:\n",
    "        ax=plt.gca()\n",
    "\n",
    "    t = ax.transData\n",
    "    fig = plt.gcf()\n",
    "\n",
    "    #horizontal version\n",
    "    for s,c in zip(ls,lc):\n",
    "        text = plt.text(x,y,\" \"+s,color=c, transform=t, **kw)\n",
    "        text.draw(fig.canvas.get_renderer())\n",
    "        ex = text.get_window_extent()\n",
    "        t = transforms.offset_copy(text._transform, y=-yoffset*ex.height, units='dots')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "id": "a467ed21-6bb6-4feb-8944-c8748adf4932",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEaCAYAAAA7YdFPAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMywgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/MnkTPAAAACXBIWXMAAAsTAAALEwEAmpwYAAA9T0lEQVR4nO3deXgUVdb48e/JDiGsAgJBFmWHsL4IKhBExwUEQXyBEQVcUHRcRkdHR0d01N/rKDqIyszgoCg6IKIijuC4IYKCCLIIkU32RQhrwhKSTs7vj6qETtLdadJJOiHn8zz9pPtW3apTlaRO171Vt0RVMcYYY4orItwBGGOMqdgskRhjjAmJJRJjjDEhsURijDEmJJZIjDHGhMQSiTHGmJBYIjGVhoisE5FkP9OSRWRXgLrTROTpUorraRE5ICK/lsbyjSltlkjMWUFEtonIZQXKRovI4tzPqtpOVb8u8+ACEJHGwANAW1U9N8RlBUyGpUFE3haRvSKSJiIbReTWsly/KR8skRgTXk2Ag6q6P9yBiEhUMar9H9BUVasDA4GnRaRryUZmyjtLJKbS8D5rEZEqbnPVYRFJAf6nwLydReRHEUkXkXeBuALTB4jIKhE5IiLfiUhSgfX8QUTWiMhREXlXRPLVd+e7DPgcaCgix0Rkmlvew13mERFZ7d0cJyJjRORnN64tInK7Wx4PzPda1jERaViwSa7gWYsb6x9FZA1wXESiAq2/IFVdp6qncj+6r/P9/hLMWckSiamsxuMc8M4HrgBG5U4QkRhgDjAdqA28B1znNb0L8DpwO1AH+CcwV0RivZb/v8CVQDMgCRhdMABV/QK4CtijqtVUdbSINAI+AZ521/0H4H0RqetW2w8MAKoDY4C/iUgXVT1eYFnVVHVPkPtiBNAfqAnUL2L9hYjIZBE5AawH9gLzglyvOUtYIjFnkznut+gjInIEmBxg3v8FnlHVQ6q6E5jkNa0HEA1MVNUsVZ0N/OA1/Tbgn6r6vapmq+qbwCm3Xq5JqrpHVQ8BHwOdgtyGkcA8VZ2nqjmq+jmwHLgaQFU/UdVf1LEQ+AzoFeSy/ZmkqjtV9WRR6/dFVe8EEtw4PsDZF6YSsURizibXqmrN3BdwZ4B5GwI7vT5vLzBtt+Yf0dR7ehPggQJJq7FbL5f3FVgngGpBbkMT4PoCy74EaAAgIleJyFIROeROuxo4J8hl++O9HwKu3x83oS4GEoFxIcZjKpjidK4ZczbYi3PwX+d+Pq/AtEYiIl7J5DzgF/f9TpyzmWdKIa6dwHRVva3gBLfp7H3gJuAjVc0SkTmAuLP4Gsr7OFDV67OvK8O86/ldf5CisD6SSsfOSExlNQt4RERqiUgicLfXtCWAB7jH7XweAnT3mv4acIeIXCiOeBHpLyIJJRDX28A1InKFiESKSJzbQZ4IxACxQCrgEZGrgN941d0H1BGRGl5lq4CrRaS2iJwL3BfC+vMRkXoiMlxEqrnzXoHT3/JVMbfdVFCWSExl9SROc9VWnH6G6bkTVDUTGILTQX4YGIbT9p87fTlOP8kr7vTN+OhMLw63v2YQ8CechLETeBCIUNV04B6cJHgY+C0w16vuemAGsMVtlmrobtdqYJu7ne8Wd/2+ZsdpxtrlxjMBuE9VPyrGppsKTOzBVsYYY0JhZyTGGGNCYonEGGNMSEotkYhIYxFZ4N6Fu05E7nXLnxCR3e5dwatE5GqvOo+IyGYR2eB23OWWdxWRn9xpk0RE3PJY967hzSLyvYg0La3tMcYY41tpnpF4gAdUtQ3OjVp3iUhbd9rfVLWT+5oH4E4bDrTDuSN4sohEuvP/HRgLtHBfV7rltwCHVfUC4G/AX0txe4wxxvhQaveRqOpenOvxUdV0EfkZaBSgyiBgpjtuz1YR2Qx0F5FtQHVVXQIgIm8B1+KMKzQIeMKtPxt4pcC1/4Wcc8452rRp0xC2zBhjKp8VK1YcUFWfQ+WUyQ2JbpNTZ+B74GLgdyJyE87QCw+o6mGcJLPUq9outyzLfV+wHPfnTgBV9YjIUZyxjw4UWP9YnDMazjvvPJYvX16Sm2eMMWc9Ednub1qpd7aLSDWcu3HvU9U0nGaq83HGHtoLvJA7q4/qGqA8UJ38BapTVLWbqnarW9fv2HPGGGOKoVQTiYhE4ySRd1T1AwBV3eeOy5ODc4dw7h3Du3CGrMiVCOxxyxN9lOerI86zFGoAh0pna4wxxvhSmldtCTAV+FlVX/Qq9x78bTCw1n0/FxjuXonVDKdTfZnb15LuPiNBcMcZ8qqTO/z3UOCrQP0jxhhjSl5p9pFcDNwI/CQiq9yyPwEjRKQTThPUNpxnOqCq60RkFpCCc8XXXaqa7dYbB0wDquB0ss93y6cC092O+UM4V30Zc1bJyspi165dZGRkhDsUUwnExcWRmJhIdHR00HUq3RAp3bp1U+tsNxXJ1q1bSUhIoE6dOri3UBlTKlSVgwcPkp6eTrNmzfJNE5EVqtrNVz27s92Yci4jI8OSiCkTIkKdOnXO+OzXnkcSLjkKzy6GmnFwp/u48HX74b0UUIWLGsMVF/iuO301/LQfEmLgz33KLmZvX26B79znITWqDjcmQbR7/+hjX0FcFESI83r4ksBxZ2XDi0vAk+Psl84NYEDL/Os7dBLeXAVpp5xlXnweXJr/G9PZzJKIKSvF+VuzRBIuC7bCudUgw+N8zlF4dx3cc6GTXP66GJLqQwMfj7jokQh9mjoH1nA4kgFfb3OSQUwk/OtHWL4HenpddHdfD6gWk7+ev7ijIuDeHk7yyc6BF5ZAu7rQrNbpeSIFrmsL59Vw9tmzi6HNOb73z9nszk9Kb9mT+5fess1ZzRJJUf62BH7bAepXg2OZzudQzwIOn4S1++HKC+DLrU7ZtiNQtyqc4z7MrmtDWL3P94GyRR04eCL49e1Kg1nrnPj3HXMuc7i6ReFv/WciW50ziUiBzGyoEVd0HX9xizhJJHe52TmF56kRd3odcVFOEj6SUaETiciTPstVx5dxJMaExhJJUVJPQN145/3uNKcZp6AXvoNT2YXLh7SB1j4epz07BQa3OX02As5BsVaV059rxTnJJVRZ2TD1RxjVCZrWhI83QFYO9G9ReN5gt6NmHFzW3GnCio50zgzaet3oKcDL3zvvezWBS86jSLlNfanHoXeT/GcjBR08ATuPOttjytwTTzxBtWrV+MMf/uB3ntTUVAYMGEBmZiaTJk2iV69eZ7SOadOmsXz5cl555RXmzJlDy5Ytadu2rc95J06cSO3atbnpppvOaB0Fff3110yYMIH//Oc/QdeZOHEiY8eOpWrVqn7nCWZ/laXLLruM9957j1q1AvyPnSFLJIEcPOEcNCPcNsPd6dDQxzfgBy4Kfpk/7XOafM6rARsPlkycgaw/AI1rnD7oNqru9MX4agcNdjtOZMGaffCXvlA1Gl77Eb7fBRcmnl5OzThIPwWTvof68c7ZSCARAn/q5Sz7n8thj599neGBKStgaFuoEvzliWelkmiKKqWmsi+//JLWrVvz5ptvhrysOXPmMGDAAJ+JxOPx8Prrr/Pjjz+GvJ7imDhxIiNHjgyYSMqCx+MhKiq4w/mNN97I5MmTefTRR0ts/XbVViC70qCR18Fsx9H8n3O98B38v0WFX+sPFJ73l8NOh/NjX8HrK2HDAXhjpXPgPXzy9HyHM4JrLipKwQPyjqNOEvMl2O1YfwDqVIGEWIiMgE7nwpbDp6fXdONOiIWO557ZmVXVaGhZx0l2BWXnwGsroHsjp0PelJlnnnmGVq1acdlll7Fhw4a88l9++YUrr7ySrl270qtXL9avX8+qVat46KGHmDdvHp06deLkyZOMGzeObt260a5dO8aPP91017RpUw4ccP6+li9fTnJycr71fvfdd8ydO5cHH3yQTp068csvv+Sb/tVXX9GlS5e8g2hycjJ//OMf6d69Oy1btmTRokWAc+XbmDFj6NChA507d2bBggU+tzMtLY3BgwfTtm1b7rjjDnJynGZWX/FPmjSJPXv20LdvX/r27QvAp59+SpcuXejYsSP9+vXLW25KSgrJyck0b96cSZMm+Vx3tWrVePTRR+nYsSM9evRg3759AGzfvp1+/fqRlJREv3792LFjBwCjR4/m/vvvp2/fvvzxj39k9OjRjBs3jr59+9K8eXMWLlzIzTffTJs2bRg9enTeegYOHMiMGTN8xlBsqlqpXl27dtWgfbJR9Y2Vzvt9x1R//6nqwRPB1y/KhgOqry5z3nuyVR/7UjX1uGpWturTC1V3p/mve+C46l++zl82cYnq4ZP5yxZtV536o/P+13TV8QtU00+FFveWQ866T3lUc3JUp61U/WqLMy0jS/Vk1un3zy1WXbsvcNxpGarHM533pzyqE75VXfNr/nlycpzfxay1ocVejsATPl8FpaSknP4w7j+nXyUhiOUtX75c27dvr8ePH9ejR4/q+eefr88//7yqql566aW6ceNGVVVdunSp9u3bV1VV33jjDb3rrrvylnHw4EFVVfV4PNqnTx9dvXq1qqo2adJEU1NTVVX1hx9+0D59+hSqP2rUKH3vvfd8xvb444/rpEmT8j736dNH77//flVV/eSTT7Rfv36qqjphwgQdPXq0qqr+/PPP2rhxYz15Mv//yoIFCzQ2NlZ/+eUX9Xg8etlll+WtN5j49+/fr4mJibply5Z8dcaPH689e/bUjIwMTU1N1dq1a2tmZmahbQF07ty5qqr64IMP6lNPPaWqqgMGDNBp06apqurUqVN10KBBefulf//+6vF48j4PGzZMc3JydM6cOZqQkKBr1qzR7Oxs7dKli65cuTJvXRdccIEeOHDA5z5VLfA3dzq+5ernuGpNW4HsPOpclfTMN06T0LnVYOkup6O6pEVGwLD28Moyp7+gZ2L+M4lXl8ENSc63/ddXOs1ixzLhT186/R09Gzv9OVULNPd0a+g0Qz210GlSu7lz4aupzlSzWs4Zwf8tcpqkGtc43Q+Snuk0TYGzHd0aQrt6zmdfcV98Hhw9BW+tduZXdS406FA//3YfOAHLdjv75P853zIZ2Ara1wttW0yRFi1axODBg/OabwYOHAjAsWPH+O6777j++uvz5j116pTPZcyaNYspU6bg8XjYu3cvKSkpJCUlhRzb3r17adOmTb6yIUOGANC1a1e2bdsGwOLFi7n77rsBaN26NU2aNGHjxo2FYujevTvNmzcHYMSIESxevJihQ4cGFf/SpUvp3bt33o18tWvXzpvWv39/YmNjiY2NpV69euzbt4/ExMR89WNiYhgwYEBe7J9//jkAS5Ys4YMPPgCcZqmHHnoor871119PZGRk3udrrrkGEaFDhw7Ur1+fDh06ANCuXTu2bdtGp06dAKhXrx579uyhTp0impyDZIkkkD3p8Eiv01cUlbSWdZxXrvb1/B8Y7+p++v3NnQtP35PuNDHFROYvj4s6fZ9KSRrQ0vdVX+dUhUd7+67jK26AxOpO/4gvudtdM84uTy2oNC8FLsDXvQU5OTnUrFmTVatWBay7detWJkyYwA8//ECtWrUYPXp03g1vUVFRec1HxRkCpkqVKoXqxcbGAhAZGYnH41zQokGO4FFwO0UkYPzeVNXvPRi5MRWMy1t0dHRefX/zFIwxPj7e53oiIiLyrTMiIiLf8jIyMqhSpQolxfpI/Mnw5L8stbxrmOB0QBtTwnr37s2HH37IyZMnSU9P5+OPPwagevXqNGvWjPfeew9wDqSrV68uVD8tLY34+Hhq1KjBvn37mD9/ft60pk2bsmLFCgDef/99n+tPSEggPT3d57Q2bdqwefPmoLbhnXfeAWDjxo3s2LGDVq1aFZpv2bJlbN26lZycHN59910uueSSgPF7x9azZ08WLlzI1q3OJf2HDpXMQOQXXXQRM2fOBOCdd97hkksuCWl5qsqvv/5KST7gzxKJP3FR8ERyuKMwJuy6dOnCsGHD6NSpE9ddd12+S3nfeecdpk6dSseOHWnXrh0fffRRofodO3akc+fOtGvXjptvvpmLL744b9r48eO599576dWrV74mGm/Dhw/n+eefp3PnzoU626+66iq++eabIrfhzjvvJDs7mw4dOjBs2DCmTZuW7xt7rp49e/Lwww/Tvn17mjVrxuDBgwPGP3bsWK666ir69u1L3bp1mTJlCkOGDKFjx44MGzasyLiCMWnSJN544w2SkpKYPn06L730UkjLW7FiBT169Aj6Kq9g2KCNxoRJsDck/vzzz4X6AcxpgwcP5rnnnqNFi1LouzwL3XvvvQwcODDfVWUF+fqbs0EbjTFnrWeffZa9e/eGO4wKo3379gGTSHFUkA4AY4zxrVWrVj77O4xvt912W4kv0xJJGExed4jYCEEEIhBGt64JwLL9J1lz0LkapG5cJP2bJBAVUfgqkC1pmXyx6zg5qnSsE0fPc8v+rtpPtqfzS1omVaMiuLXN6aEW0jKz+c/2YxzPykEEOtaJ43/qOVeHZHhymL/zGKknsxHg6ibVaBQfHbBOMNvqLxZjTNmwRBImI1rUoGrU6ZbF9MxsVqSe5NY2tYiOEOZsTSPl8CmS6uS/uz1Hlc92HmP4BTVIiI5g2oYjtKgRwzlVyvZX2aFOHF3rVuE/2/NfTRMhwqWN4jm3ahSnsnOYtuEIzRKiOadKFF/sPk7zhBgGN4sjO0fJytGAdWrHRQa1rf5iMcaUDUskRXhn0xGubFyNOnFRnPTk8M6mo6X2rTdHwZOjRApk5SgJ0YW7sPae8FArNpKasc4VLm1rxbLpaGbARLLvhIfPdx3jZLZyMMMZlPHic6vQq0G83zpFOa9aNEd8DPBYLTqCam7csZER1ImLIj0rh4SYHHYey6L/edUAiIwQIt2zLX91TuVoUNvqLxZjTNmwRFKEI6dyqO0eyPaf9FA3rvAlim9vPEJmTuGr3y5tGE/T6oXvIhfg3c1HAeh8ThU6nRNHQkwk3etVYfK6Q0RFCM0SYmjmo256Zg4JMacTTEJMBHuO+75xCZzE9NG2dAY0qUbD+Gi+2XMcj8IlPpqIznQ7inLkVDb7T3hoGB/FkVM5VI2K4JMdx9h/0sO5VaO4rFE1YiLFb52taVlntK2VxbMrfYzhFoKHO/sYodrLtm3bGDBgAGvXrg16mevXr2f48OGICLNnz+b8888PNcw81apV49ixY0HPX9TowVC8kX/LSnkbPdgXSyQBHM3Mplp0RN6dpPtPZlPPxzf/kS1rntFyR7asQUJ0JMezcpi5+Si14yKpFxfJpqOZjGtbm9goYc7WdNYeyqB97dAGbtyWnsW5VaNoGO8MnVKvShRb0jJ93oF7ptsRSGa28uHWNPolxhMbGUEO2fx6wsPlifE0jI/m813HWLrvBL0bxvutYyquOXPmMGjQIJ580vclzgXljtkUEVHyv/dAoweXpTMZobcslURc5W+rypF9JzzUq3L6DOTXEx7a1Cp8E9OZfpNPcB9JGx8dQcuaMew9nsWJrBxqxkRQ1W3iaVkjht3HPbSvXaBuTATpmacf/JSemeOzCSxXaoGzqF9Peqhf1fevvaTOSLLVSQjtasfRqqazvxKiI0mIichLaK1rxrJ034nAdc5wW03p8Xg8jBo1ipUrV9KyZUveeustqlatyooVK7j//vs5duwY55xzDtOmTWPlypVMnDiRyMhIvvnmGxYsWMCLL77I66+/DsCtt97Kfffdx7Zt2/Ju5luyZAlz5sxh1qxZzJo1i1OnTjF48GC/ieiBBx5gwYIF1KpVi5kzZ1K3bl1ee+01pkyZQmZmJhdccAHTp09n1apVzJ07l4ULF/L000/z/vvvo6rccccdpKamEhkZmXdn/rFjxxg6dChr166la9euvP3224W+cCUnJ3PhhReyYMECjhw5wtSpU+nVqxcZGRmMGzeO5cuXExUVxYsvvkjfvn2ZNm0an3zyCRkZGRw/fpybbrqJOXPmkJ2dzdq1a3nggQfIzMxk+vTpxMbGMm/evHxjdBXkaxuzs7NJSkpi48aNREdHk5aWRlJSEps2bWLHjh3cddddpKamUrVqVV577TVat27N6NGjqV27NitXrqRLly688MILIf192H9lAPtPZuNxj2OHMrLZfDSTulUKN22NbFmTm1vXKvTydfDNzFZOuU8AzMxWtqVnUbdKFNVjIthzwkNWjvPNbPuxLOr4aEZrUDWKQ6eyOXIqm+wcJeXwKS6o4axnxqajpGfm7yuoEhXB/pOevG3YeCSTtj6S4Zluhz+qyrztx6gT5zTV5aoWHUH16AgOug/z2paeSR13+Bl/dQJtqylbGzZsYOzYsaxZs4bq1aszefJksrKyuPvuu5k9ezYrVqzg5ptv5tFHH+Xqq6/mjjvu4Pe//z0LFixgxYoVvPHGG3z//fcsXbqU1157jZUrV+Yt96abbmLlypVs2LCBTZs2sWzZMlatWsWKFSt83rV+/PhxunTpwo8//kifPn3yks2QIUP44YcfWL16NW3atGHq1KlcdNFFDBw4kOeff55Vq1Zx/vnnc8MNN3DXXXexevVqvvvuOxo0cB5JkJsAU1JS2LJlC99++63PfeHxeFi2bBkTJ07MW/err74KwE8//cSMGTMYNWpU3nhcS5Ys4c033+Srr74CYO3atfz73/9m2bJlPProo1StWpWVK1fSs2dP3nrrrYC/B1/bmJCQQHJyMp984oy9NnPmTK677jqio6MZO3YsL7/8MitWrGDChAnceeedecvauHEjX3zxRchJBOyMJKD9Jz1ERQhTfz5MvSpR1ImLZO2hU1wcwuW2Jzw5vL8lDXCeeNu2VizN3QN1q5oxvLH+CBEC9atE0cnriq1ZvxzlqvOqkRAdyW8Sq/HuL0dRhaQ6cdStEoWqcvhUNnFR+b8bOB3Up/jXz4epEiUMbJpAlajQvz98tDWNHceyOOlRXl17iEsaVKVjnTh2Hfew7vAp6sZF8vp65xklfRrEc36NGC5PrMbH246RrUrN2Mi8jvdAdXxta8H94S8WU3IaN26cNzTIyJEjmTRpEldeeSVr167l8ssvByA7OzvvoOxt8eLFDB48OG+AwSFDhrBo0SIGDhxIkyZN6NGjBwCfffYZn332GZ07O4N7Hjt2jE2bNtG7d/5BQCMiIvKGHxk5cmTeaL9r167lscce48iRIxw7dowrrriiUCzp6ens3r2bwYMHAxAXd/rvpHv37nkj8nbq1Ilt27b5HNfqTEYXBrj88svznWX07duXhIQEEhISqFGjBtdccw0AHTp0YM2aNYXW583fNt56660899xzXHvttbzxxhu89tprRY7OXHDk4FBYIglg/0kPY1rXLNH2+pqxkdzi56qvXg3i/V5J9b/nn34Y1fk1Yji/Rv7T3wMZ2bSqGUN0gftOYiKF673qlpRBzXw8chhoXC3ab+dt/apReffMBFvH17ZC/v3hLxZTcnyNiquqtGvXjiVLlgSsG2gYJu/Ra1WVRx55hNtvv71YsY0ePZo5c+bQsWNHpk2bxtdff31GsQQzQq/3fMGOLuxvhF7IP0pvwRF6ffG3jRdffDHbtm1j4cKFZGdn0759e9LS0gKOzlwwrlBY05Yfp7Kdm+MqSqdv3SpR9EusFu4wzFlqx44deQljxowZXHLJJbRq1YrU1NS88qysLNatW1eobu/evZkzZw4nTpzg+PHjfPjhhz6f4X7FFVfw+uuv512RtXv3bvbvL/ykzJycHGbPng3Av//977yzhvT0dBo0aEBWVlbeSL+Qf4Te6tWrk5iYyJw5cwDnG/qJEycIVbCjC4fK3zYC3HTTTYwYMYIxY8YAwY/OXBLsjMSP2MgIbm/rv9PLmHAp6nLd0tCmTRvefPNNbr/9dlq0aMG4ceOIiYlh9uzZ3HPPPRw9ehSPx8N9991Hu3bt8tXt0qULo0ePpnt359kyt956K507d85rFsr1m9/8hp9//pmePXsCzmW+b7/9NvXq5X9GT3x8POvWraNr167UqFGDd999F4CnnnqKCy+8kCZNmtChQ4e85DF8+HBuu+02Jk2axOzZs5k+fTq33347jz/+ONHR0XkH2lDceeed3HHHHXTo0IGoqCi/owuHyt82Atxwww089thjjBgxIq/snXfeYdy4cTz99NNkZWUxfPhwOnbsWOJx2ei/xoSJjf5rStLs2bP56KOPmD59esjLOtPRf+2MxBhjKri7776b+fPnM2/evLCs3xKJMcZUcC+//HJY118xepLPYp9+uplWrV7hggsm8eyzi0Oer7TdfPNH1Kv3PO3bTw44X3Z2Dp07/5MBA/5dRpEZY8LFEkkYZWfncNdd85g//wZSUu5ixoy1pKSkFnu+sjB6dCc+/XRkkfO99NL3tGlT9p3CxpiyZ4mkCMnJ09iwwRkk7+DBE0V+Ez8Ty5bt5oILatO8eS1iYiIZPrwdH320vtjzFbR69a/07v0Gbdu+SkTEk4g8yfjxC0KKuXfvJtSuXSXgPLt2pfHJJ5u49dYuIa3LGFMxWB9JETZvPkSLFnUAWLNmHx061C80T69eb5CefqpQ+YQJv+Gyy5r7Xfbu3ek0bnz6ZrrExOp8//3uYs/nLSPDw7Bhs3nrrcF0796IP//5KzIyPDzxRHKJxB7Iffd9ynPPXUZ6emax6hv//F3pVRIKXi1mTLAskQSwffsRGjWqToR7t/iaNftISqpXaL5Fi8YUa/m+Lr32MShv0PN5++KLLXTp0oDu3RsBkJRUn08/3VzoDuXixu7Pf/6zkXr14unatSFff72tRJdtyp/33nuPxx9/nHPPPZfx48cTExPDRRddVOzlFWc494kTJzJ27FiqVvU/dFF5Hoo9OTmZCRMm0K2bzytrK4RSSyQi0hh4CzgXyAGmqOpLIlIbeBdoCmwD/ldVD7t1HgFuAbKBe1T1v255V2AaUAWYB9yrqioise46ugIHgWGquq2ktmHVql/zJY4VK/YybFi7QvOdybf6V19dxmuv/ei+v5qdO9Pypu3alUbDhgmFlpOYWD2o+bytXbufDh1Ox/7jj3vp0qXwOEglfUby7bc7mDt3A/PmbSIjw0Na2ilGjvyAt98ecsbLMuXf1KlTmTx5Mn379s07WJ9JIimJIcwnTpzIyJEjAyaSslBeh4nPzs4usTG1/Mp9DkBJv4AGQBf3fQKwEWgLPAc87JY/DPzVfd8WWA3EAs2AX4BId9oyoCfOM6HmA1e55XcC/3DfDwfeLSqurl27arCefPJrHTnyA1VV3bjxgFav/n+6ffuRoOsXJSsrW5s1m6hbthzSU6c8mpT0d127dt8ZzXfppW/qrl1HC9WZMmW5Dh8+W1VVN2w4oC1bvqwHDhwvkbi3bj2s7dq9WuR8CxZs1f793ymRdZ6N4Amfr4JSUlJ81inpGAIZNGiQdunSRdu2bav//Oc/VVX1ySef1Pj4eG3ZsqUOHTpU69evrw0bNtSOHTvqN998o/v379chQ4Zot27dtFu3brp48WJVVR0/frzedtttevnll+uIESPyrWfBggXaq1cvvfbaa7VNmzZ6++23a3Z2tqqq3nHHHdq1a1dt27atPv7446qq+tJLL2l0dLS2b99ek5OTVVV1/vz52rlzZ01KStJLL700b51jxozRPn36aLNmzfSll17yuZ3x8fH6pz/9SZOSkvTCCy/UX3/9VVVVt23bppdeeql26NBBL730Ut2+fbuqqo4aNUp///vfa3Jyst5///06atQoveOOOzQ5OVmbNWumX3/9tY4ZM0Zbt26to0aN8rnOPn366A8//OB3G7/44gu99tpr8+b/7LPPdPDgwaqq+t///ld79OihnTt31qFDh2p6erqqqjZp0kSffPJJvfjii3XGjBkBf7e+eP/N5QKWq7/jvb8JJf0CPgIuBzYADfR0stngvn8EeMRr/v+6yaMBsN6rfATwT+953PdRwAHcu/X9vc4kkQwePFN/+9v3NSnp7zpy5Ad64YWv6V/+8nXQ9YPxyScbtUWLSdq8+Uv69NML80276qq3dffuNL/zZWfn6Hnn/U1PnMgstNz09FM6YMC/tV27V7VPnzd0xYo9JRLv8OGz9dxzJ2hU1F+0UaMX9F//WlEo1lyWSAKrSInk4MGDqqp64sQJbdeunR44cEBV8x8Ex48fr88//3xenREjRuiiRYtUVXX79u3aunXrvPm6dOmiJ06cKLSeBQsWaGxsrP7yyy/q8Xj0sssu0/feey9fDB6PR/v06aOrV69WVeegmZqaqqqq+/fv18TERN2yZUu+OuPHj9eePXtqRkaGpqamau3atTUzs/D/DaBz585VVdUHH3xQn3rqKVVVHTBggE6bNk1VVadOnaqDBg1SVSeR9O/fXz0eT97nYcOGaU5Ojs6ZM0cTEhJ0zZo1mp2drV26dNGVK1cWWqf3PvS1jTk5OdqqVSvdv39/3n6dO3eupqamaq9evfTYsWOqqvrss8/qk08+mbdP/vrXvxZaV7DONJGUyXmYiDQFOgPfA/VVdS+Aqu4Vkdz2l0bAUq9qu9yyLPd9wfLcOjvdZXlE5ChQByehhGzNmn2sXHk7CQklP2ZOrquvbsHVV7fwOW3evBsCzpeSksp117WhSpXoQnWrVYvh449HFCoP1YwZ1/ks9441V3JyU5KTm5Z4DKbsTZo0iQ8//BCAnTt3smnTJurUqROwzhdffEFKSkre57S0tLyxoQYOHEiVKr6v/uvevTvNmzvNqiNGjGDx4sUMHTqUWbNmMWXKFDweD3v37iUlJYWkpKR8dZcuXUrv3r1p1qwZQL7h2/v3709sbCyxsbHUq1ePffv25Q0bnysmJoYBAwYAzjDxn3/+OeA8U+SDDz4A4MYbb+Shhx7Kq1NwOPZrrrkGEaFDhw7Ur1+fDh06ANCuXTu2bdtGp06d/O4zf9t444038vbbbzNmzBiWLFnCW2+9xaeffkpKSkre8P6ZmZl545QBeUPtl4VSTyQiUg14H7hPVdN8PeI1d1YfZRqgPFCdgjGMBcYCnHfeeUWFDEB6+ikiIqRUk0io2revx4svFn7mgjEl6euvv+aLL75gyZIlVK1aleTk5LyHNgWSk5PDkiVLfCaMQEOY+xqyfuvWrUyYMIEffviBWrVqMXr0aJ8xqKrPx0hDcMPER0dH59UPNJS89zr8DRPvPUR87udAw8QH2sYxY8ZwzTXXEBcXx/XXX09UlPMMossvv5wZM2b4XF5JDhNflFK9j0REonGSyDuq+oFbvE9EGrjTGwC540TvAhp7VU8E9rjliT7K89URkSigBnCoYByqOkVVu6lqt7p16wYVe0JCLBs33h3UvMaEg8iTIb+CcfToUWrVqkXVqlVZv349S5cu9Tmf93Dt4Izm+8orr+R99vdcjIKWLVvG1q1bycnJ4d133+WSSy4hLS2N+Ph4atSowb59+5g/f77P9fbs2ZOFCxeydetWAA4dKnQ4KJaLLrqImTNnAs6Iur4eeBWqQNvYsGFDGjZsyNNPP83o0aMB6NGjB99++y2bN28G4MSJE3kP0yprpZZIxEnZU4GfVfVFr0lzgVHu+1E4fSe55cNFJFZEmgEtgGVuM1i6iPRwl3lTgTq5yxoKfOW25RljSsiVV16Jx+MhKSmJP//5z3lPNCzommuu4cMPP6RTp04sWrSISZMmsXz5cpKSkmjbti3/+Mc/glpfz549efjhh2nfvj3NmjVj8ODBdOzYkc6dO9OuXTtuvvnmvOYcgLFjx+Y9+71u3bpMmTKFIUOG0LFjxxJr3pk0aRJvvPEGSUlJTJ8+nZdeeqlElust0DaCM0x848aNadu2LQB169Zl2rRpjBgxgqSkJHr06MH69UXfqFwaSm0YeRG5BFgE/IRz+S/An3D6SWYB5wE7gOtV9ZBb51HgZsCD0xQ23y3vxunLf+cDd6uqikgcMB2n/+UQMFxVtwSKy4aRN+VFcYaRtxsSK6/f/e53dO7cmVtuuaXU11VuhpFX1cX47sMA6OenzjPAMz7KlwPtfZRnANcXLDfmbGUH+8qpa9euxMfH88ILL4Q7FJ/K390zxhhj8lmxYkW4QwjIBm00pgKwrj9TVorzt2aJxJhyLi4ujoMHD1oyMaVOVTl48CBxcXFnVM+atowp5xITE9m1axepqeF5Bo2pXOLi4grdqFkUSyTGlHPR0dF5d2obUx5Z05YxxpiQWCIxxhgTEkskxhhjQmKJxBhjTEgskRhjjAmJJRJjjDEhsURijDEmJJZIjDHGhMQSiTHGmJBYIjHGGBMSSyTGGGNCYonEGGNMSCyRGGOMCYklEmOMMSGxRGKMMSYklkiMMcaExBKJMcaYkFgiMcYYExJLJMYYY0JiicQYY0xILJEYY4wJiSUSY4wxIbFEYowxJiSWSIwxxoTEEokxxpiQWCIxxhgTEkskxhhjQmKJxBhjTEgskRhjjAmJJRJjjDEhKbVEIiKvi8h+EVnrVfaEiOwWkVXu62qvaY+IyGYR2SAiV3iVdxWRn9xpk0RE3PJYEXnXLf9eRJqW1rYYY4zxrzTPSKYBV/oo/5uqdnJf8wBEpC0wHGjn1pksIpHu/H8HxgIt3FfuMm8BDqvqBcDfgL+W1oYYY4zxr9QSiap+AxwKcvZBwExVPaWqW4HNQHcRaQBUV9UlqqrAW8C1XnXedN/PBvrlnq0YY4wpO+HoI/mdiKxxm75quWWNgJ1e8+xyyxq57wuW56ujqh7gKFDH1wpFZKyILBeR5ampqSW3JcYYY8o8kfwdOB/oBOwFXnDLfZ1JaIDyQHUKF6pOUdVuqtqtbt26ZxSwMcaYwMo0kajqPlXNVtUc4DWguztpF9DYa9ZEYI9bnuijPF8dEYkCahB8U5oxxpgSUqaJxO3zyDUYyL2iay4w3L0SqxlOp/oyVd0LpItID7f/4ybgI686o9z3Q4Gv3H4UY4wxZSiqtBYsIjOAZOAcEdkFjAeSRaQTThPUNuB2AFVdJyKzgBTAA9ylqtnuosbhXAFWBZjvvgCmAtNFZDPOmcjw0toWY4wx/pVaIlHVET6KpwaY/xngGR/ly4H2PsozgOtDidEYY0zogmraEpF4EYlw37cUkYEiEl26oRljjKkIgu0j+QaIE5FGwJfAGJzmJmOMMZVcsIlEVPUEMAR4WVUHA21LLyxjjDEVRdCJRER6AjcAn7hlpda/YowxpuIINpHcBzwCfOheYdUcWFBqURljjKkwgjqrUNWFwEIRiXc/bwHuKc3AjDHGVAzBXrXVU0RSgJ/dzx1FZHKpRmaMMaZCCLZpayJwBXAQQFVXA71LKSZjjDEVSNBDpKjqzgJF2T5nNMYYU6kEe+XVThG5CFARicHpH/m59MIyxhhTUQR7RnIHcBennw/Syf1sjDGmkgv2qq0DOPeQGGOMMfkETCQi8pCqPiciL+PjoVGqapcAG2NMJVfUGUluP8jy0g7EGGNMxRQwkajqxyISCbRX1QfLKCZjjDEVSJGd7e4DprqWQSzGGGMqoGAv/10pInOB94DjuYWq+kGpRGWMMabCCDaR1Ma5q/1SrzIFLJEYY0wlF2wi+ZeqfutdICIXl0I8xhhjKphgb0h8OcgyY4wxlUxR95H0BC4C6orI/V6TqgORpRmYMcaYiqGopq0YoJo7X4JXeRowtLSCMsYYU3EUdR9J7gOtpqnq9jKKyRhjTAUSbGd7rIhMAZp611HVS/3WMMYYUykEm0jeA/4B/At7DokxxhgvwSYSj6r+vVQjMcYYUyEFe/nvxyJyp4g0EJHaua9SjcwYY0yFEOwZySj3p/fAjQo0L9lwjDHGVDTBPtiqWWkHYowxpmIK2LQlIg95vb++wLT/V1pBGWOMqTiK6iMZ7vX+kQLTrizhWIwxxlRARSUS8fPe12djjDGVUFGJRP289/XZGGNMJVRUZ3tHEUnDOfuo4r7H/RxXqpEZY4ypEAKekahqpKpWV9UEVY1y3+d+jg5UV0ReF5H9IrLWq6y2iHwuIpvcn7W8pj0iIptFZIOIXOFV3lVEfnKnTRIRcctjReRdt/x7EWla7L1gjDGm2IK9IbE4plG4Q/5h4EtVbQF86X5GRNridOy3c+tMFpHcYer/DowFWriv3GXeAhxW1QuAvwF/LbUtMcYY41epJRJV/QY4VKB4EPCm+/5N4Fqv8pmqekpVtwKbge4i0gCorqpLVFWBtwrUyV3WbKBf7tmKMcaYslOaZyS+1FfVvQDuz3pueSNgp9d8u9yyRu77guX56qiqBzgK1PG1UhEZKyLLRWR5ampqCW2KMcYYKPtE4o+vMwkNUB6oTuFC1Smq2k1Vu9WtW7eYIRpjjPGlrBPJPre5Cvfnfrd8F9DYa75EYI9bnuijPF8dEYkCalC4Kc0YY0wpK+tEMpfTA0COAj7yKh/uXonVDKdTfZnb/JUuIj3c/o+bCtTJXdZQ4Cu3H8UYY0wZCnb03zMmIjOAZOAcEdkFjAeeBWaJyC3ADuB6AFVdJyKzgBTAA9ylqrkP0BqHcwVYFWC++wKYCkwXkc04ZyLew7kYY4wpI6WWSFR1hJ9J/fzM/wzwjI/y5UB7H+UZuInIGGNM+JSXznZjjDEVlCUSY4wxIbFEYowxJiSWSIwxxoTEEokxxpiQWCIxxhgTEkskxhhjQmKJxBhjTEgskRhjjAmJJRJjjDEhsURijDEmJJZIjDHGhMQSiTHGmJBYIjHGGBMSSyTGGGNCYonEGGNMSCyRGGOMCYklEmOMMSGxRGKMMSYklkiMMcaExBKJMcaYkFgiMcYYExJLJMYYY0JiicQYY0xILJEYY4wJiSUSY4wxIYkKdwDGmPxEnsz3WXV8mCIxJjh2RmKMMSYklkiMMcaExJq2jCkHCjZfFWzeMqY8szMSY4wxIbFEYowxJiSWSIwxxoTEEokxxpiQhCWRiMg2EflJRFaJyHK3rLaIfC4im9yftbzmf0RENovIBhG5wqu8q7uczSIySUQkHNtjjDGVWTjPSPqqaidV7eZ+fhj4UlVbAF+6nxGRtsBwoB1wJTBZRCLdOn8HxgIt3NeVZRi/McYYylfT1iDgTff9m8C1XuUzVfWUqm4FNgPdRaQBUF1Vl6iqAm951THGGFNGwpVIFPhMRFaIyFi3rL6q7gVwf9ZzyxsBO73q7nLLGrnvC5YXIiJjRWS5iCxPTU0twc0wxhgTrhsSL1bVPSJSD/hcRNYHmNdXv4cGKC9cqDoFmALQrVs3n/MYY4wpnrCckajqHvfnfuBDoDuwz22uwv253519F9DYq3oisMctT/RRbowxpgyVeSIRkXgRSch9D/wGWAvMBUa5s40CPnLfzwWGi0isiDTD6VRf5jZ/pYtID/dqrZu86hhjjCkj4Wjaqg986F6pGwX8W1U/FZEfgFkicguwA7geQFXXicgsIAXwAHepara7rHHANKAKMN99GWOMKUNlnkhUdQvQ0Uf5QaCfnzrPAM/4KF8OtC/pGI0xxgSvPF3+a4wxpgKyRGKMMSYklkiMMcaExBKJMcaYkNgTEo0pQ/bkQ3M2sjMSY4wxIbFEYowxJiTWtGVMmKiOD3cIxpQIOyMxxhgTEkskxhhjQmKJxBhjTEgskRhjjAmJdbYbU85533tiHfSmPLIzEmOMMSGxRGKMMSYk1rRlTDnk3YRlw6qY8s7OSIwxxoTEEokxxpiQWNOWMaXImqVMZWBnJMYYY0JiicQYY0xIrGnLmDJSEjcTFmwqsxsUTXlgZyTGGGNCYonEGGNMSKxpy5hyrmDzlV0JZsobSyTGlDA70JvKxpq2jDHGhMTOSIypwGyIeVMeWCIxpgT4a86yg7upDCyRGHOWsLMTEy6WSIwphnB2qNsQ86a8sURiTAkrD2cDdge8KUuWSIwJUnn89m9nJ6Y8sERiTADBHJwrwrd96z8xpckSiTFeKvK3+mDvgA+0jZZkTHFU+EQiIlcCLwGRwL9U9dkwh2TKidJMCmfrATfYfXa2br8pngqdSEQkEngVuBzYBfwgInNVNaVMArjzkzOvM7l/cMsIdr5A9YKsI39fHtyyKzkd1+30h+L87suYd7wl/TuuyGduocr3d1DaSvN4Mbl/4OWdwd+4qGrQM5c3ItITeEJVr3A/PwKgqv/nr07DOufr3kM3lVGEpiIr0wNGGNkXCROcJ1aoqs9/igp9RgI0AnZ6fd4FXFhwJhEZC4x1Px6DJzaUQWxFOQc4EO4gyolyuS/k72W+ynK5H8LE9sVp5WVfNPE3oaInEvFRVugUS1WnAFNKP5zgichyf9m9srF94bD9cJrti9Mqwr6o6KP/7gIae31OBPaEKRZjjKmUKnoi+QFoISLNRCQGGA7MDXNMxhhTqVTopi1V9YjI74D/4lz++7qqrgtzWMEqV01tYWb7wmH74TTbF6eV+31Roa/aMsYYE34VvWnLGGNMmFkiMcYYExJLJGEmIn8QERWRc8IdS7iIyPMisl5E1ojIhyJSM9wxlTURuVJENojIZhF5ONzxhIuINBaRBSLys4isE5F7wx1TuIlIpIisFJH/hDsWfyyRhJGINMYZ3mVHuGMJs8+B9qqaBGwEHglzPGXKa6ifq4C2wAgRaRveqMLGAzygqm2AHsBdlXhf5LoX+DncQQRiiSS8/gY8hI+bKCsTVf1MVT3ux6U49wNVJt2Bzaq6RVUzgZnAoDDHFBaquldVf3Tfp+McQBuFN6rwEZFEoD/wr3DHEoglkjARkYHAblVdHe5YypmbgfnhDqKM+Rrqp9IePHOJSFOgM/B9mEMJp4k4XzZzwhxHQBX6PpLyTkS+AM71MelR4E/Ab8o2ovAJtC9U9SN3nkdxmjbeKcvYyoGghvqpTESkGvA+cJ+qpoU7nnAQkQHAflVdISLJYQ4nIEskpUhVL/NVLiIdgGbAahEBpynnRxHprqq/lmGIZcbfvsglIqOAAUA/rXw3N9lQP15EJBonibyjqh+EO54wuhgYKCJXA3FAdRF5W1VHhjmuQuyGxHJARLYB3VS1PIzwWebch5O9CPRR1dRwx1PWRCQK5yKDfsBunKF/fluBRmkoMeJ8s3oTOKSq94U5nHLDPSP5g6oOCHMoPlkfiSkPXgESgM9FZJWI/CPcAZUl90KD3KF+fgZmVcYk4roYuBG41P1bWOV+IzflmJ2RGGOMCYmdkRhjjAmJJRJjjDEhsURijDEmJJZIjDHGhMQSiTHGmJBYIjHGGBMSSyTGGGNCYomkkhKRwe5zUFoXMV9NEbmzQNl3xVxniS0rFCJyj/u8i3d8xVTMZR4ridjOYH2vi8h+EVlbzPp5+6BAeb79ISJNi7uOAsudWhL7ucAyS+R352O5JbLNlYklksprBLAYGF7EfDWBfP+sqnpRMddZkssKxZ3A1ap6g6+YiiKOMvvf8bO+acCVISzWex94q8kZ7o9ARGSQiGwBBgKPicgPItKuhBZfkxKMFfKGaLHj4plSVXtVshdQDWdMp5bAeq/ym4A1wGpguls2EzgJrAKed8uOuT//CtzpVf8J4AFgDrACWAeM9ZoeaFn3A2vd131uWVOcIUNec5f1GVDFnRYPfOLGuhYY5mM7C8UB/APIBH4Cfu8nppHAMrfsn0CkVyyTgZVAkwLrOlbEep8C7vWa5xngnuKuz2v/rC3id+1rv+bbBwXmz7c/ivgdFIq7wLLOxxl8siXOMDijgSHu8grO63c9AfZRwVgf8tqnfwO+ct/3A94O4u8sd1/3yd2vQHO37H/C/X9bnl9hD8BeYfilO/+UU9333wFdgHbABuAct7y2+7PQwYrTB//OwEKv8hTgPK+6Vdx/2DqBlgV0dQ9q8ThJbp277KY4w8p3cuedBYx0318HvOa1nBo+ttNfHNu8tjNfTEAb4GMg2v08GSfBNsV5JkQPP/v0WKD1uvV/dMsjgF/c8mKtz9/+LDDd534tuA8CLdPf78Bf3AWWNQ6Y7L5/BRjtvl8HtPaxXn+/60D7yDvWHsB77vtFOIknGhgP3O5vfxTc17nLBVrhJJFO4f6fLe8vG0a+chqB88AccL7VjcB53O9sdUcgVtVDRS1EVVeKSD0RaQjUBQ6r6g4ReUJEBruzNQZaAAcDLOoS4ENVPQ4gIh8AvYC5wFZVXeXOtwLnnxycA8IEEfkr8B9VXeRjufecYRzgfHvtCvzgDvFfBdgPfANsV9WlRdT3uV5VXSoiB0WkM1AfWKmqB0VkRAmszx9/+3XlGS7H1++gpp+4C/I1mJ/g+xks/n7XgX4n3lYAXUUkATgF/Ah0w9nme3Aea+3v76zgvq4LfARcp5V3AM2gWSKpZESkDnAp0F5EFKeJQIEXKN7DlGYDQ3EeWjXTHe76MqCnqp4Qka9xnqUQMKwA0055vc/GOYigqhtFpCtwNfB/IvKZqv4lb4HFiyM3ljdVNd9z492n9R0vsnLg9f4Lp3nnXOD1klhfUeGEWD+Xr9+Bz7gL+AL4s4i8mBeQyCCc487GINeDv3W5+yiPqma5j2QYg3OmvQboi9PE9jOBHyRXcF8fxXlq5cU4Zy4mAOtUqnyGAm+pahNVbaqqjYGtOO3M/+smGkSktjt/Os4Q7/7MxOmwH4qTVGrgnJmccK8I6+E1r79lfQNcKyJVRSQeGIzTNOGXexZ0QlXfBibgNM95CxSHt4IxfQkMFZF67npqi0iTQLGcwXo/xOkg/x+cIeNLYn2BnPF+pejfd64i41bVTTjD438JXI/TT/QXnG/52WewHf7W5SvWb4A/uD8XAXcAq9RpszqT/ZEJXAvcJCK/PYNYKyVLJJXPCJwDmrf3cZLBM8BCEVmN86ApVPUg8K2IrBWR5wsuzD3tT8B5/vxe4FMgSkTW4Bw4lnrN63NZqvojzlVIy3Cez/0vVS2q+aUDsExEVuE8uvjpAtP9xlEg/nwxqWoK8BjwmVv3c6BBEbEEtV5VzQQW4DxvJNstK9b6RGQGsARoJSK7ROQWH9t2xvu1qN+313xBxa3OEw6bAfOA51S1o6r+VNT2BbMuP7EucuNYoqr7gAy37Iz3h9sENgD4vXsmZfyw55EYU0bcS3h/BK53v60bc1awMxJjyoCItAU2A19aEjFnGzsjMcYYExI7IzHGGBMSSyTGGGNCYonEGGNMSCyRGGOMCYklEmOMMSGxRGKMMSYklkiMMcaE5P8DqDyPe6LOA0kAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for feat in range(6):\n",
    "\n",
    "    nbins=100\n",
    "    r=(-5,5)\n",
    "\n",
    "    plt.hist(act[:,:,feat][tmask],nbins,r,color='hotpink',lw=3,label='default (no batch norm)',histtype='step')\n",
    "    \n",
    "    plt.hist(bnIn[:,:,feat][tmask],nbins,r,color='skyblue',label='before batch norm layer')\n",
    "    plt.hist(bnOut[:,:,feat][tmask],nbins,r,color='navy',lw=3,\n",
    "             label='after batch norm layer',histtype='step')\n",
    "    \n",
    "    '''\n",
    "    Add text to the plot to track the mean + std\n",
    "    '''\n",
    "    lc = ['hotpink','skyblue','navy']\n",
    "    ls = []\n",
    "    for o in [act,bnIn,bnOut]:\n",
    "        oi = o[:,:,feat][tmask]\n",
    "    \n",
    "        ls.append(f'$\\mu$ = {oi.mean():.1f}, $\\sigma$ = {oi.std():.1f}')\n",
    "    \n",
    "    _, ymax = plt.ylim()\n",
    "    rainbow_text(-4.8,.98*ymax,ls,lc, ha='left',va='top',yoffset=1)\n",
    "    \n",
    "    plt.xlim(r)\n",
    "    plt.legend(loc='upper right')\n",
    "    plt.xlabel(f'Activations after layer {phi_i+1} of the $\\Phi$ network')\n",
    "    plt.ylabel('Entries')\n",
    "    plt.title(f'Hidden feature {feat}')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b616e238-cc56-4d0b-a93c-77aa18353b22",
   "metadata": {},
   "source": [
    "**Also look at activations in the F network**"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9ac233a4-cc80-4d49-8c53-5dfcbaacee3e",
   "metadata": {},
   "source": [
    "Now that we have the branches for the (pieces) of the network setup - evaluate them to get the corresponding activations."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "id": "d4924e67-990f-4dbe-89aa-7246c6e6ce07",
   "metadata": {},
   "outputs": [],
   "source": [
    "ni = 100000\n",
    "\n",
    "F_j = 0"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "id": "e903cc95-f63a-4510-9373-a3efac114d7b",
   "metadata": {},
   "outputs": [],
   "source": [
    "func = K.function([dips.get_layer('Input').input], dips.get_layer(f'F{F_j}_Dense').output)\n",
    "act = func([X_test[:ni]])  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "id": "1fc98f46-16cb-4954-bb8a-12e1d3d7e65f",
   "metadata": {},
   "outputs": [],
   "source": [
    "'''\n",
    "TO DO - fill in to get the activations fro the batch norm network\n",
    "'''\n",
    "\n",
    "func = K.function([dips_bn.get_layer('Input').input], \n",
    "                  dips_bn.get_layer(f'F{F_j}_Dense').output)\n",
    "bnIn = func([X_test[:ni]])  \n",
    "\n",
    "func = K.function([dips_bn.get_layer('Input').input], \n",
    "                  dips_bn.get_layer(f'F{F_j}_BatchNormalization').output)\n",
    "bnOut = func([X_test[:ni]])  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d3149fef-6c29-4f29-85ff-d23057fea24b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "id": "f743d35d-232c-4060-88f9-a3356a9a1cf1",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for feat in range(6):\n",
    "\n",
    "    nbins=100\n",
    "    r=(-5,5)\n",
    "\n",
    "    plt.hist(act[:,feat],nbins,r,color='hotpink',lw=3,\n",
    "             label='default (no batch norm)',histtype='step')\n",
    "    \n",
    "    plt.hist(bnIn[:,feat],nbins,r,color='skyblue',label='before batch norm layer')\n",
    "    plt.hist(bnOut[:,feat],nbins,r,color='navy',lw=3,\n",
    "             label='after batch norm layer',histtype='step')\n",
    "    \n",
    "    '''\n",
    "    Add text to the plot to track the mean + std\n",
    "    '''\n",
    "    lc = ['hotpink','skyblue','navy']\n",
    "    ls = []\n",
    "    for o in [act,bnIn,bnOut]:\n",
    "        oi = o[:,feat]\n",
    "        ls.append(f'$\\mu$ = {oi.mean():.1f}, $\\sigma$ = {oi.std():.1f}')\n",
    "    \n",
    "    _, ymax = plt.ylim()\n",
    "    rainbow_text(-4.8,.98*ymax,ls,lc, ha='left',va='top',yoffset=1)\n",
    "    \n",
    "    plt.xlim(r)\n",
    "    plt.legend(loc='upper right')\n",
    "    plt.xlabel(f'Activations after layer {F_j+1} of the $F$ network')\n",
    "    plt.ylabel('Entries')\n",
    "    plt.title(f'Hidden feature {feat}')\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5ae13c68-328f-463b-ae47-17102ca6b5b6",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "df46bb26-1a4d-432b-be18-fafdeed48f3f",
   "metadata": {},
   "source": [
    "And... that's it for today!!\n",
    "\n",
    "\n",
    "Just as a reminder what we learned in this tutorial:\n",
    "\n",
    "1. How to setup and train a deep sets model for a multi-class classification problem\n",
    "2. To boost our understanding - we tested some other permutation invariant operations\n",
    "3. We checked the impact of batch norm on the models and examined the activations\n",
    "\n",
    "But the biggest take away is... hopefully you're convinced that Deep Sets is a fast and easy model to get running - and inspired to test it out on some of your ATLAS research problems :) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d820b6f8-5cbb-47f5-8567-4afe458f42e8",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0054bfcb",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9f0e3d5f",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}