LeNet_MNIST_Optimization_Experiment.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.7.6"
    },
    "colab": {
      "name": "LeNet_MNIST-Optimization_Experiment.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "accelerator": "TPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "oc_q-vSYIwoL",
        "colab_type": "text"
      },
      "source": [
        "# Lottery Ticket Hypothesis: _LeNet_ Dense Fully-Connected Feed Forward Neural Network for MNIST\n",
        "\n",
        "Multi-class classification of MNIST dataset by following the implementation of the hypothesis using the following architecture:\n",
        "\n",
        "__Fully Connected Layers:__ 300, 100, 10\n",
        "\n",
        "This code is optimized where the weights of the trained neural network are _not_ saved to disk, but rather used in a sequential manner, while training using the _iterative pruning rounds_.\n",
        "\n",
        "This will reduce the time taken to:\n",
        "\n",
        "1. Save weights on disk\n",
        "1. Load weights from disk"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "R3scUag1I1j3",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "8bfa3ef2-8299-4e5a-b983-0de433c7c64c"
      },
      "source": [
        "%tensorflow_version 2.x"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "TensorFlow 2.x selected.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OII2KxjHIwoT",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import math\n",
        "import tensorflow_model_optimization as tfmot\n",
        "from tensorflow_model_optimization.sparsity import keras as sparsity\n",
        "# from tensorflow.keras import datasets, layers, models\n",
        "import matplotlib.pyplot as plt\n",
        "from tensorflow.keras.layers import AveragePooling2D, Conv2D, MaxPooling2D, ReLU\n",
        "from tensorflow.keras import models, layers, datasets\n",
        "from tensorflow.keras.layers import Dense, Flatten, Reshape, Input, InputLayer\n",
        "from tensorflow.keras.models import Sequential, Model\n",
        "# from tensorflow.keras.initializers import RandomNormal\n",
        "# import math\n",
        "from sklearn.metrics import accuracy_score, precision_score, recall_score"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RFNDOrRaIwom",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "f8c6e4fe-68b6-4075-d708-08c583f8ba5a"
      },
      "source": [
        "tf.__version__"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'2.1.0'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AMria2VnIwo1",
        "colab_type": "code",
        "colab": {},
        "outputId": "8a9af62d-be08-4f7d-8fe3-3f74dad78b22"
      },
      "source": [
        "%env CUDA_DEVICE_ORDER=PCI_BUS_ID\n",
        "%env CUDA_VISIBLE_DEVICES=2"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "env: CUDA_DEVICE_ORDER=PCI_BUS_ID\n",
            "env: CUDA_VISIBLE_DEVICES=2\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gEi8K75_IwpA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "batch_size = 60\n",
        "num_classes = 10\n",
        "num_epochs = 100"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d7tgalWEIwpO",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        },
        "outputId": "7a1aa13d-646d-4851-c570-e88457260d47"
      },
      "source": [
        "# Data preprocessing and cleadning:\n",
        "# input image dimensions\n",
        "img_rows, img_cols = 28, 28\n",
        "\n",
        "# Load MNIST dataset-\n",
        "(X_train, y_train), (X_test, y_test) = tf.keras.datasets.mnist.load_data()"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
            "11493376/11490434 [==============================] - 0s 0us/step\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jMvyvwOgIwpZ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "outputId": "d534daa0-48f0-49be-841b-a99a3919e0c0"
      },
      "source": [
        "if tf.keras.backend.image_data_format() == 'channels_first':\n",
        "    X_train = X_train.reshape(X_train.shape[0], 1, img_rows, img_cols)\n",
        "    X_test = X_test.reshape(X_test.shape[0], 1, img_rows, img_cols)\n",
        "    input_shape = (1, img_rows, img_cols)\n",
        "else:\n",
        "    X_train = X_train.reshape(X_train.shape[0], img_rows, img_cols, 1)\n",
        "    X_test = X_test.reshape(X_test.shape[0], img_rows, img_cols, 1)\n",
        "    input_shape = (img_rows, img_cols, 1)\n",
        "\n",
        "print(\"\\n'input_shape' which will be used = {0}\\n\".format(input_shape))"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "'input_shape' which will be used = (28, 28, 1)\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_XgxWo8zIwph",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Convert datasets to floating point types-\n",
        "X_train = X_train.astype('float32')\n",
        "X_test = X_test.astype('float32')\n",
        "\n",
        "# Normalize the training and testing datasets-\n",
        "X_train /= 255.0\n",
        "X_test /= 255.0"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1QkXJLllIwpq",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# convert class vectors/target to binary class matrices or one-hot encoded values-\n",
        "y_train = tf.keras.utils.to_categorical(y_train, num_classes)\n",
        "y_test = tf.keras.utils.to_categorical(y_test, num_classes)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "afPACTGVIwp0",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Reshape training and testing sets-\n",
        "X_train = X_train.reshape(X_train.shape[0], 784)\n",
        "X_test = X_test.reshape(X_test.shape[0], 784)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EyVIuGDmIwp-",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86
        },
        "outputId": "a4537063-f5be-4749-96eb-0f959344e6c2"
      },
      "source": [
        "print(\"\\nDimensions of training and testing sets are:\")\n",
        "print(\"X_train.shape = {0}, y_train.shape = {1}\".format(X_train.shape, y_train.shape))\n",
        "print(\"X_test.shape = {0}, y_test.shape = {1}\".format(X_test.shape, y_test.shape))"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Dimensions of training and testing sets are:\n",
            "X_train.shape = (60000, 784), y_train.shape = (60000, 10)\n",
            "X_test.shape = (10000, 784), y_test.shape = (10000, 10)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hVAgRgoMIwqJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tcb-v8r8IwqS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VN8U5n-DIwqb",
        "colab_type": "text"
      },
      "source": [
        "### Prepare MNIST dataset for _GradientTape_ training:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "UJXCrvB8Iwqg",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Create training and testing datasets-\n",
        "train_dataset = tf.data.Dataset.from_tensor_slices((X_train, y_train))\n",
        "test_dataset = tf.data.Dataset.from_tensor_slices((X_test, y_test))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qN69v2j-Iwqo",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "train_dataset = train_dataset.shuffle(buffer_size = 20000, reshuffle_each_iteration = True).batch(batch_size = batch_size, drop_remainder = False)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Uj8ly5IHIwqx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "test_dataset = test_dataset.batch(batch_size=batch_size, drop_remainder=False)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3x3thjbdIwq5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Choose an optimizer and loss function for training-\n",
        "loss_fn = tf.keras.losses.CategoricalCrossentropy()\n",
        "optimizer = tf.keras.optimizers.Adam(lr = 0.0012)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mPiwLvOKIwrA",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Select metrics to measure the error & accuracy of model.\n",
        "# These metrics accumulate the values over epochs and then\n",
        "# print the overall result-\n",
        "train_loss = tf.keras.metrics.Mean(name = 'train_loss')\n",
        "train_accuracy = tf.keras.metrics.CategoricalAccuracy(name = 'train_accuracy')\n",
        "\n",
        "test_loss = tf.keras.metrics.Mean(name = 'test_loss')\n",
        "test_accuracy = tf.keras.metrics.CategoricalAccuracy(name = 'test_accuracy')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cZsRaagaIwrJ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lG_caIjYIwrS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ND2k8VZXIwrZ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "67818ee7-1ec1-44ea-ea0d-5946ee4cd3d9"
      },
      "source": [
        "# The model is first trained without any pruning for 'num_epochs' epochs-\n",
        "epochs = num_epochs\n",
        "\n",
        "num_train_samples = X_train.shape[0]\n",
        "\n",
        "end_step = np.ceil(1.0 * num_train_samples / batch_size).astype(np.int32) * epochs\n",
        "\n",
        "print(\"'end_step parameter' for this dataset =  {0}\".format(end_step))"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "'end_step parameter' for this dataset =  100000\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZVlZb33mIwrj",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Specify the parameters to be used for layer-wise pruning, NO PRUNING is done here:\n",
        "pruning_params_unpruned = {\n",
        "    'pruning_schedule': sparsity.ConstantSparsity(\n",
        "        target_sparsity=0.0, begin_step=0,\n",
        "        end_step = end_step, frequency=100\n",
        "    )\n",
        "}"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ays34e0oIwrr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zWFkh9kWIwry",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "l = tf.keras.layers"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KV-xYRWdIwr6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xfozKePcIwr_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def pruned_nn(pruning_params):\n",
        "    \"\"\"\n",
        "    Function to define the architecture of a neural network model\n",
        "    following 300 100 architecture for MNIST dataset and using\n",
        "    provided parameter which are used to prune the model.\n",
        "    \n",
        "    Input: 'pruning_params' Python 3 dictionary containing parameters which are used for pruning\n",
        "    Output: Returns designed and compiled neural network model\n",
        "    \"\"\"\n",
        "    \n",
        "    pruned_model = Sequential()\n",
        "    pruned_model.add(l.InputLayer(input_shape=(784, )))\n",
        "    pruned_model.add(Flatten())\n",
        "    pruned_model.add(sparsity.prune_low_magnitude(\n",
        "        Dense(units = 300, activation='relu', kernel_initializer=tf.initializers.GlorotUniform()),\n",
        "        **pruning_params))\n",
        "    # pruned_model.add(l.Dropout(0.2))\n",
        "    pruned_model.add(sparsity.prune_low_magnitude(\n",
        "        Dense(units = 100, activation='relu', kernel_initializer=tf.initializers.GlorotUniform()),\n",
        "        **pruning_params))\n",
        "    # pruned_model.add(l.Dropout(0.1))\n",
        "    pruned_model.add(sparsity.prune_low_magnitude(\n",
        "        Dense(units = num_classes, activation='softmax'),\n",
        "        **pruning_params))\n",
        "    \n",
        "    # Compile pruned CNN-\n",
        "    pruned_model.compile(\n",
        "        loss=tf.keras.losses.categorical_crossentropy,\n",
        "        # optimizer='adam',\n",
        "        optimizer=tf.keras.optimizers.Adam(lr = 0.0012),\n",
        "        metrics=['accuracy'])\n",
        "    \n",
        "    return pruned_model\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Qj5hyhd9IwsE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nDOWeP1iIwsL",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Add a pruning step callback to peg the pruning step to the optimizer's\n",
        "# step. Also add a callback to add pruning summaries to tensorboard\n",
        "callback = [\n",
        "             sparsity.UpdatePruningStep(),\n",
        "             # sparsity.PruningSummaries(log_dir = logdir, profile_batch=0),\n",
        "             tf.keras.callbacks.EarlyStopping(\n",
        "                 monitor='val_loss', patience = 3,\n",
        "                 min_delta=0.001\n",
        "             )\n",
        "]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TrkP-XUhIwsP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wupE7JytIwsk",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        },
        "outputId": "60e5e3f1-ff26-4996-a503-5b4fac86cd8a"
      },
      "source": [
        "# Initialize a CNN model-\n",
        "orig_model = pruned_nn(pruning_params_unpruned)"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow_model_optimization/python/core/sparsity/keras/pruning_wrapper.py:183: Layer.add_variable (from tensorflow.python.keras.engine.base_layer) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Please use `layer.add_weight` method instead.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cjO7k1maIwss",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Strip model of it's pruning parameters\n",
        "orig_model_stripped = sparsity.strip_pruning(orig_model)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LArs9L-FIwsy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Save random weights-\n",
        "orig_model.save_weights(\"LeNet_MNIST_Random_Weights.h5\", overwrite=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xbnfR3wzIws5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Save random weights-\n",
        "# orig_model.save_weights(\"LeNet_MNIST_Winning_Ticket.h5\", overwrite=True)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AIu6z1HxIws9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "outputId": "00b2ac3e-2851-4e1c-95ea-439738aeb3b2"
      },
      "source": [
        "# Get CNN summary-\n",
        "orig_model_stripped.summary()"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "flatten (Flatten)            (None, 784)               0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 300)               235500    \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 100)               30100     \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 10)                1010      \n",
            "=================================================================\n",
            "Total params: 266,610\n",
            "Trainable params: 266,610\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tlVLmrRlIwtD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VhY6atf7IwtH",
        "colab_type": "text"
      },
      "source": [
        "### Layer-wise pruning:\n",
        "The Lottery Ticket Hypothesis follows a _layer-wise pruning_ heuristic, where the fully-connected (fc) layers are pruned at the rate of 20% per iterative\n",
        "pruning round.\n",
        "\n",
        "Therefore, there is a need to compute the percentage of weights being pruned for each layer."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PiloUG9GIwtI",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "outputId": "a320d992-f7bb-4bf5-aa4e-0181c7f9df03"
      },
      "source": [
        "# number of fully-connected dense parameters-\n",
        "dense1 = 235500\n",
        "dense2 = 30100\n",
        "op_layer = 1010\n",
        "\n",
        "\n",
        "# total number of parameters-\n",
        "total_params = dense1 + dense2 + op_layer\n",
        "\n",
        "print(\"\\nTotal number of trainable parameters = {0}\\n\".format(total_params))"
      ],
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Total number of trainable parameters = 266610\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "idTP01B5IwtO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# maximum pruning performed is till 0.5% of all parameters-\n",
        "max_pruned_params = 0.005 * total_params"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p27ZlhebIwtR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "loc_tot_params = total_params\n",
        "loc_dense1 = dense1\n",
        "loc_dense2 = dense2\n",
        "loc_op_layer = op_layer\n",
        "\n",
        "# variable to count number of pruning rounds-\n",
        "n = 0"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RtYl-QsvIwtX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Lists to hold percentage of weights pruned in each round for all layers in CNN-\n",
        "dense1_pruning = []\n",
        "dense2_pruning = []\n",
        "op_layer_pruning = []\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-OtVfhV5Iwta",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "548ab0fc-1d7a-4ca2-f506-88464db7dca0"
      },
      "source": [
        "while loc_tot_params >= max_pruned_params:\n",
        "    loc_dense1 *= 0.8   # 20% weights are pruned\n",
        "    loc_dense2 *= 0.8   # 20% weights are pruned\n",
        "    loc_op_layer *= 0.8 # 20% weights are pruned\n",
        "    \n",
        "    dense1_pruning.append(((dense1 - loc_dense1) / dense1) * 100)\n",
        "    dense2_pruning.append(((dense2 - loc_dense2) / dense2) * 100)\n",
        "    op_layer_pruning.append(((op_layer - loc_op_layer) / op_layer) * 100)\n",
        "\n",
        "    loc_tot_params = loc_dense1 + loc_dense2 + loc_op_layer\n",
        "\n",
        "    n += 1\n",
        "\n",
        "    print(\"Dense1 = {0:.3f}, Dense2 = {1:.3f} & O/p layer = {2:.3f}\".format(\n",
        "        loc_dense1, loc_dense2, loc_op_layer))\n",
        "    print(\"Total number of parameters = {0:.3f}\\n\".format(loc_tot_params))\n",
        "    "
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Dense1 = 188400.000, Dense2 = 24080.000 & O/p layer = 808.000\n",
            "Total number of parameters = 213288.000\n",
            "\n",
            "Dense1 = 150720.000, Dense2 = 19264.000 & O/p layer = 646.400\n",
            "Total number of parameters = 170630.400\n",
            "\n",
            "Dense1 = 120576.000, Dense2 = 15411.200 & O/p layer = 517.120\n",
            "Total number of parameters = 136504.320\n",
            "\n",
            "Dense1 = 96460.800, Dense2 = 12328.960 & O/p layer = 413.696\n",
            "Total number of parameters = 109203.456\n",
            "\n",
            "Dense1 = 77168.640, Dense2 = 9863.168 & O/p layer = 330.957\n",
            "Total number of parameters = 87362.765\n",
            "\n",
            "Dense1 = 61734.912, Dense2 = 7890.534 & O/p layer = 264.765\n",
            "Total number of parameters = 69890.212\n",
            "\n",
            "Dense1 = 49387.930, Dense2 = 6312.428 & O/p layer = 211.812\n",
            "Total number of parameters = 55912.169\n",
            "\n",
            "Dense1 = 39510.344, Dense2 = 5049.942 & O/p layer = 169.450\n",
            "Total number of parameters = 44729.736\n",
            "\n",
            "Dense1 = 31608.275, Dense2 = 4039.954 & O/p layer = 135.560\n",
            "Total number of parameters = 35783.788\n",
            "\n",
            "Dense1 = 25286.620, Dense2 = 3231.963 & O/p layer = 108.448\n",
            "Total number of parameters = 28627.031\n",
            "\n",
            "Dense1 = 20229.296, Dense2 = 2585.570 & O/p layer = 86.758\n",
            "Total number of parameters = 22901.625\n",
            "\n",
            "Dense1 = 16183.437, Dense2 = 2068.456 & O/p layer = 69.407\n",
            "Total number of parameters = 18321.300\n",
            "\n",
            "Dense1 = 12946.749, Dense2 = 1654.765 & O/p layer = 55.525\n",
            "Total number of parameters = 14657.040\n",
            "\n",
            "Dense1 = 10357.400, Dense2 = 1323.812 & O/p layer = 44.420\n",
            "Total number of parameters = 11725.632\n",
            "\n",
            "Dense1 = 8285.920, Dense2 = 1059.050 & O/p layer = 35.536\n",
            "Total number of parameters = 9380.505\n",
            "\n",
            "Dense1 = 6628.736, Dense2 = 847.240 & O/p layer = 28.429\n",
            "Total number of parameters = 7504.404\n",
            "\n",
            "Dense1 = 5302.989, Dense2 = 677.792 & O/p layer = 22.743\n",
            "Total number of parameters = 6003.523\n",
            "\n",
            "Dense1 = 4242.391, Dense2 = 542.233 & O/p layer = 18.195\n",
            "Total number of parameters = 4802.819\n",
            "\n",
            "Dense1 = 3393.913, Dense2 = 433.787 & O/p layer = 14.556\n",
            "Total number of parameters = 3842.255\n",
            "\n",
            "Dense1 = 2715.130, Dense2 = 347.029 & O/p layer = 11.645\n",
            "Total number of parameters = 3073.804\n",
            "\n",
            "Dense1 = 2172.104, Dense2 = 277.623 & O/p layer = 9.316\n",
            "Total number of parameters = 2459.043\n",
            "\n",
            "Dense1 = 1737.683, Dense2 = 222.099 & O/p layer = 7.452\n",
            "Total number of parameters = 1967.235\n",
            "\n",
            "Dense1 = 1390.147, Dense2 = 177.679 & O/p layer = 5.962\n",
            "Total number of parameters = 1573.788\n",
            "\n",
            "Dense1 = 1112.117, Dense2 = 142.143 & O/p layer = 4.770\n",
            "Total number of parameters = 1259.030\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CM0yUpMgIwtg",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86
        },
        "outputId": "75b495b5-308c-41d5-f530-2ec091b733bc"
      },
      "source": [
        "print(\"\\nnumber of pruning rounds = {0}\\n\\n\".format(n))"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "number of pruning rounds = 24\n",
            "\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mlknOX0OIwtj",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "num_pruning_rounds = n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ga6nUT4RIwts",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1XX4XOfrIwtx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Convert from list to np.array-\n",
        "dense1_pruning = np.array(dense1_pruning)\n",
        "dense2_pruning = np.array(dense2_pruning)\n",
        "op_layer_pruning = np.array(op_layer_pruning)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uSQAKBCSIwt3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Round off numpy arrays to 3 decimal digits-\n",
        "dense1_pruning = np.round(dense1_pruning, decimals=3)\n",
        "dense2_pruning = np.round(dense2_pruning, decimals=3)\n",
        "op_layer_pruning = np.round(op_layer_pruning, decimals=3)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AyvbdnrDIwt7",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "outputId": "94f1a9d7-9e41-4228-d193-29b227b674a7"
      },
      "source": [
        "dense1_pruning"
      ],
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([20.   , 36.   , 48.8  , 59.04 , 67.232, 73.786, 79.028, 83.223,\n",
              "       86.578, 89.263, 91.41 , 93.128, 94.502, 95.602, 96.482, 97.185,\n",
              "       97.748, 98.199, 98.559, 98.847, 99.078, 99.262, 99.41 , 99.528])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 36
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KprBtCp3IwuB",
        "colab_type": "text"
      },
      "source": [
        "### Note:\n",
        "\n",
        "For _tensorflow-model-optimization_ package to work, the pruning rates should be between 0 and 1, hence convert them"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SCLvUVAgIwuD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "dense1_pruning = dense1_pruning / 100\n",
        "dense2_pruning = dense2_pruning / 100\n",
        "op_layer_pruning = op_layer_pruning / 100\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BoxnAV7vIwuJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 86
        },
        "outputId": "6f9d98ac-f46e-498b-d0e9-00e553aff709"
      },
      "source": [
        "dense1_pruning"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([0.2    , 0.36   , 0.488  , 0.5904 , 0.67232, 0.73786, 0.79028,\n",
              "       0.83223, 0.86578, 0.89263, 0.9141 , 0.93128, 0.94502, 0.95602,\n",
              "       0.96482, 0.97185, 0.97748, 0.98199, 0.98559, 0.98847, 0.99078,\n",
              "       0.99262, 0.9941 , 0.99528])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 38
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bhGfvFJsIwuO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9iQuerFjIwuS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "D_a9U-v6IwuX",
        "colab_type": "text"
      },
      "source": [
        "### Create a mask:\n",
        "A mask is created where all parameters equal to one.\n",
        "\n",
        "This will be used later by _GradientTape_ for training the defined neural network."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kldJJlWDIwuZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Instantiate a new neural network model for which, the mask is to be created,\n",
        "# according to the paper-\n",
        "mask_model = pruned_nn(pruning_params_unpruned)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6jVx_odGIwuc",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Strip the model of its pruning parameters-\n",
        "mask_model_stripped = sparsity.strip_pruning(mask_model)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wbFaaDCDIwuh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Assign all masks to one-\n",
        "\n",
        "for wts in mask_model_stripped.trainable_weights:\n",
        "    wts.assign(\n",
        "        tf.ones_like(\n",
        "            input = wts,\n",
        "            dtype = tf.float32\n",
        "        )\n",
        "\n",
        "    )\n",
        "    # wts.assign(1.)\n",
        "    # wts.assign(tf.where(tf.equal(wts, 0.), 0., 1.))"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8bMozWg0Iwul",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 243
        },
        "outputId": "299da626-9140-4ecd-b621-1f37a343ca9d"
      },
      "source": [
        "print(\"\\nMask model metrics:\")\n",
        "print(\"layer-wise number of nonzero parameters in each layer are: \\n\")\n",
        "\n",
        "masked_sum_params = 0\n",
        "\n",
        "for layer in mask_model_stripped.trainable_weights:\n",
        "    print(tf.math.count_nonzero(layer, axis = None).numpy())\n",
        "    masked_sum_params += tf.math.count_nonzero(layer, axis = None).numpy()\n",
        "\n",
        "print(\"\\nTotal number of trainable parameters = {0}\\n\".format(masked_sum_params))"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Mask model metrics:\n",
            "layer-wise number of nonzero parameters in each layer are: \n",
            "\n",
            "235200\n",
            "300\n",
            "30000\n",
            "100\n",
            "1000\n",
            "10\n",
            "\n",
            "Total number of trainable parameters = 266610\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "VuDWceRtIwup",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "P9swbfX9Iwus",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U2DgEFZIIwuw",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "outputId": "7620b008-1b40-4c4a-e87f-f307f41aad3f"
      },
      "source": [
        "print(\"\\nnumber of pruning rounds for LeNet NN = {0} and number of epochs = {1}\\n\".format(num_pruning_rounds, num_epochs))"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "number of pruning rounds for LeNet NN = 24 and number of epochs = 100\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NwGhUzDgIwu1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S7E5a4GKIwu5",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ghAWZVe4IwvC",
        "colab_type": "text"
      },
      "source": [
        "### Python 3 dictionary to hold model training metrics for each of _n_ rounds:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "StajMbeaIwvD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Method - 1: Nested Python 3 dictionaries:\n",
        "history_main = {}\n",
        "\n",
        "# for x in range(num_pruning_rounds + 1):\n",
        "for x in range(num_pruning_rounds):\n",
        "    history = {}\n",
        "    \n",
        "    # Neural Network model, scalar metrics-\n",
        "    history['accuracy'] = np.zeros(shape = num_epochs)\n",
        "    history['val_accuracy'] = np.zeros(shape = num_epochs)\n",
        "    history['loss'] = np.zeros(shape = num_epochs)\n",
        "    history['val_loss'] = np.zeros(shape = num_epochs)\n",
        "    \n",
        "    # compute % of weights pruned at the end of each iterative pruning round-\n",
        "    history['percentage_wts_pruned'] = 90\n",
        "\n",
        "    history_main[x + 1] = history\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZLyojYkNIwvH",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "24dba34b-f7c7-4afb-817f-d3866557e39a"
      },
      "source": [
        "# Sanity check-\n",
        "history_main.keys()"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "dict_keys([1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wv8qGKoKIwvM",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "5b849b81-bc13-426f-d127-6d5f6e073944"
      },
      "source": [
        "# Sanity check-\n",
        "history_main[10]['accuracy'].shape"
      ],
      "execution_count": 46,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(100,)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 46
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1u2l-jKVIwvQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4_mYz3rnIwvS",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# User input parameters for Early Stopping in manual implementation-\n",
        "minimum_delta = 0.001\n",
        "patience = 3"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7uaicTPKIwvX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "best_val_loss = 100\n",
        "loc_patience = 0"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "naHBb_DLIwva",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "orig_sum_params = 266610"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "w-tMeu0VIwvd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sns8zCq1Iwvg",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ytsZoBHjXP6c",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Instantiate a model-\n",
        "winning_ticket_model = pruned_nn(pruning_params_unpruned)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o5K8E0N1XXjs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Use Randomly initialized weights-\n",
        "winning_ticket_model.set_weights(orig_model.get_weights())"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-8C_U0mDXell",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CSnCEsFtXfGs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lcVracjVIwvk",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "outputId": "1a3fe483-ad0c-44c6-ddf1-8896fa22d882"
      },
      "source": [
        "for i in range(1, num_pruning_rounds + 1):\n",
        "    \n",
        "    print(\"\\n\\n\\nIterative pruning round: {0}\\n\\n\".format(i))\n",
        "    \n",
        "    # Define 'train_one_step()' and 'test_step()' functions here-\n",
        "    @tf.function\n",
        "    def train_one_step(model, mask_model, optimizer, x, y):\n",
        "        '''\n",
        "        Function to compute one step of gradient descent optimization\n",
        "        '''\n",
        "        with tf.GradientTape() as tape:\n",
        "            # Make predictions using defined model-\n",
        "            y_pred = model(x)\n",
        "\n",
        "            # Compute loss-\n",
        "            loss = loss_fn(y, y_pred)\n",
        "        \n",
        "        # Compute gradients wrt defined loss and weights and biases-\n",
        "        grads = tape.gradient(loss, model.trainable_variables)\n",
        "    \n",
        "        # type(grads)\n",
        "        # list\n",
        "    \n",
        "        # List to hold element-wise multiplication between-\n",
        "        # computed gradient and masks-\n",
        "        grad_mask_mul = []\n",
        "    \n",
        "        # Perform element-wise multiplication between computed gradients and masks-\n",
        "        for grad_layer, mask in zip(grads, mask_model.trainable_weights):\n",
        "            grad_mask_mul.append(tf.math.multiply(grad_layer, mask))\n",
        "    \n",
        "        # Apply computed gradients to model's weights and biases-\n",
        "        optimizer.apply_gradients(zip(grad_mask_mul, model.trainable_variables))\n",
        "\n",
        "        # Compute accuracy-\n",
        "        train_loss(loss)\n",
        "        train_accuracy(y, y_pred)\n",
        "\n",
        "        return None\n",
        "    \n",
        "    \n",
        "    @tf.function\n",
        "    def test_step(model, optimizer, data, labels):\n",
        "        \"\"\"\n",
        "        Function to test model performance\n",
        "        on testing dataset\n",
        "        \"\"\"\n",
        "    \n",
        "        predictions = model(data)\n",
        "        t_loss = loss_fn(labels, predictions)\n",
        "\n",
        "        test_loss(t_loss)\n",
        "        test_accuracy(labels, predictions)\n",
        "\n",
        "        return None\n",
        "\n",
        "\n",
        "    # Instantiate a model\n",
        "    model_gt = pruned_nn(pruning_params_unpruned)\n",
        "    \n",
        "    # Load winning ticket (from above)-\n",
        "    # model_gt.load_weights(\"LeNet_MNIST_Winning_Ticket.h5\")\n",
        "    model_gt.set_weights(winning_ticket_model.get_weights())\n",
        "    \n",
        "    # Strip model of pruning parameters-\n",
        "    model_gt_stripped = sparsity.strip_pruning(model_gt)\n",
        "    \n",
        "    \n",
        "    # Train model using 'GradientTape'-\n",
        "    \n",
        "    # Initialize parameters for Early Stopping manual implementation-\n",
        "    best_val_loss = 100\n",
        "    loc_patience = 0\n",
        "    \n",
        "    for epoch in range(num_epochs):\n",
        "    \n",
        "        if loc_patience >= patience:\n",
        "            print(\"\\n'EarlyStopping' called!\\n\")\n",
        "            break\n",
        "        \n",
        "        # Reset the metrics at the start of the next epoch\n",
        "        train_loss.reset_states()\n",
        "        train_accuracy.reset_states()\n",
        "        test_loss.reset_states()\n",
        "        test_accuracy.reset_states()\n",
        "        \n",
        "        \n",
        "        for x, y in train_dataset:\n",
        "            # train_one_step(model_gt_stripped, mask_model, optimizer, x, y, grad_mask_mul)\n",
        "            train_one_step(model_gt_stripped, mask_model_stripped, optimizer, x, y)\n",
        "\n",
        "\n",
        "        for x_t, y_t in test_dataset:\n",
        "            # test_step(x_t, y_t)\n",
        "            test_step(model_gt_stripped, optimizer, x_t, y_t)\n",
        "\n",
        "        template = 'Epoch {0}, Loss: {1:.4f}, Accuracy: {2:.4f}, Test Loss: {3:.4f}, Test Accuracy: {4:4f}'\n",
        "    \n",
        "        # 'i' is the index for number of pruning rounds-\n",
        "        history_main[i]['accuracy'][epoch] = train_accuracy.result() * 100\n",
        "        history_main[i]['loss'][epoch] = train_loss.result()\n",
        "        history_main[i]['val_loss'][epoch] = test_loss.result()\n",
        "        history_main[i]['val_accuracy'][epoch] = test_accuracy.result() * 100\n",
        "        \n",
        "        print(template.format(epoch + 1, \n",
        "                              train_loss.result(), train_accuracy.result()*100,\n",
        "                              test_loss.result(), test_accuracy.result()*100))\n",
        "    \n",
        "        # Count number of non-zero parameters in each layer and in total-\n",
        "        # print(\"layer-wise manner model, number of nonzero parameters in each layer are: \\n\")\n",
        "\n",
        "        model_sum_params = 0\n",
        "    \n",
        "        for layer in model_gt_stripped.trainable_weights:\n",
        "            # print(tf.math.count_nonzero(layer, axis = None).numpy())\n",
        "            model_sum_params += tf.math.count_nonzero(layer, axis = None).numpy()\n",
        "    \n",
        "        print(\"Total number of trainable parameters = {0}\\n\".format(model_sum_params))\n",
        "\n",
        "    \n",
        "        # Code for manual Early Stopping:\n",
        "        if np.abs(test_loss.result() < best_val_loss) >= minimum_delta:\n",
        "            # update 'best_val_loss' variable to lowest loss encountered so far-\n",
        "            best_val_loss = test_loss.result()\n",
        "        \n",
        "            # reset 'loc_patience' variable-\n",
        "            loc_patience = 0\n",
        "        \n",
        "        else:  # there is no improvement in monitored metric 'val_loss'\n",
        "            loc_patience += 1  # number of epochs without any improvement\n",
        "\n",
        "    \n",
        "    # Resize numpy arrays according to the epoch when 'EarlyStopping' was called-\n",
        "    for metrics in history_main[i].keys():\n",
        "        history_main[i][metrics] = np.resize(history_main[i][metrics], new_shape = epoch)\n",
        "        # history[metrics] = np.resize(history[metrics], new_shape=epoch)\n",
        "     \n",
        "    \n",
        "    # Save trained model weights-\n",
        "    # model_gt.save_weights(\"LeNet_MNIST_Trained_Weights.h5\", overwrite=True)\n",
        "    \n",
        "    \n",
        "    \n",
        "    \n",
        "    # Prune trained model:\n",
        "    \n",
        "    # print(\"\\n% of weights to be pruned in round = {0} is: {1:.4f}\\n\".format(i, wts_np[i - 1]))\n",
        "    \n",
        "    # Specify the parameters to be used for layer-wise pruning, Fully-Connected layer pruning-\n",
        "    pruning_params_fc = {\n",
        "        'pruning_schedule': sparsity.ConstantSparsity(\n",
        "            target_sparsity=dense1_pruning[i - 1], begin_step = 1000,\n",
        "            end_step = end_step, frequency=100\n",
        "        )\n",
        "    }\n",
        "    \n",
        "    \n",
        "    # Instantiate a Nueal Network model to be pruned using parameters from above-\n",
        "    pruned_model = pruned_nn(pruning_params_fc)\n",
        "    \n",
        "    # Load weights from original trained and unpruned model-\n",
        "    # pruned_model.load_weights(\"LeNet_MNIST_Trained_Weights.h5\")\n",
        "    pruned_model.set_weights(model_gt.get_weights())\n",
        "    \n",
        "    # Train pruned NN-\n",
        "    history_pruned = pruned_model.fit(\n",
        "        x = X_train, y = y_train,\n",
        "        batch_size = batch_size,\n",
        "        epochs = epochs,\n",
        "        verbose = 1,\n",
        "        callbacks = callback,\n",
        "        validation_data = (X_test, y_test),\n",
        "        shuffle = True\n",
        "    )\n",
        "    \n",
        "    # Strip the pruning wrappers from pruned model-\n",
        "    pruned_model_stripped = sparsity.strip_pruning(pruned_model)\n",
        "    \n",
        "    # print(\"\\nIn pruned model, number of nonzero parameters in each layer are: \\n\")\n",
        "    pruned_sum_params = 0\n",
        "    \n",
        "    for layer in pruned_model_stripped.trainable_weights:\n",
        "        # print(tf.math.count_nonzero(layer, axis = None).numpy())\n",
        "        pruned_sum_params += tf.math.count_nonzero(layer, axis = None).numpy()\n",
        "    \n",
        "    print(\"\\nRound = {0}, total number of trainable parameters = {1}\\n\".format(i, pruned_sum_params))\n",
        "    # print(\"\\nTotal number of trainable parameters = {0}\\n\".format(pruned_sum_params))\n",
        "    \n",
        "    '''\n",
        "    # Sanity-check: confirm that the weights are actually pruned away from the network-\n",
        "    print(\"\\nRound = {0}, % of weights pruned away = {1:.2f}%\\n\".format( \\\n",
        "                                                i, (orig_sum_params - pruned_sum_params) / orig_sum_params * 100))\n",
        "    '''\n",
        "    \n",
        "    # 'i' is the index for number of pruning rounds-\n",
        "    history_main[i]['percentage_wts_pruned'] = ((orig_sum_params - pruned_sum_params) / orig_sum_params) * 100\n",
        "\n",
        "    # Save weights of PRUNED and Trained model BEFORE stripping-\n",
        "    # pruned_model.save_weights(\"LeNet_MNIST_Pruned_Weights.h5\", overwrite = True)\n",
        "    \n",
        "    \n",
        "    # Create a mask:\n",
        "    \n",
        "    # Instantiate a new neural network model for which, the mask is to be created,\n",
        "    mask_model = pruned_nn(pruning_params_unpruned)\n",
        "    \n",
        "    # Load weights of PRUNED model-\n",
        "    # mask_model.load_weights(\"LeNet_MNIST_Pruned_Weights.h5\")\n",
        "    mask_model.set_weights(pruned_model.get_weights())\n",
        "    \n",
        "    # Strip the model of its pruning parameters-\n",
        "    mask_model_stripped = sparsity.strip_pruning(mask_model)\n",
        "    \n",
        "    # For each layer, for each weight which is 0, leave it, as is.\n",
        "    # And for weights which survive the pruning,reinitialize it to ONE (1)-\n",
        "    for wts in mask_model_stripped.trainable_weights:\n",
        "        wts.assign(tf.where(tf.equal(wts, 0.), 0., 1.))\n",
        "    \n",
        "    \n",
        "    \n",
        "    \n",
        "    # Extract Winning Ticket:\n",
        "    \n",
        "    # Instantiate a new neural network model for which, the weights are to be extracted-\n",
        "    winning_ticket_model = pruned_nn(pruning_params_unpruned)\n",
        "    \n",
        "    # Load weights of PRUNED model-\n",
        "    # winning_ticket_model.load_weights(\"LeNet_MNIST_Pruned_Weights.h5\")\n",
        "    winning_ticket_model.set_weights(pruned_model.get_weights())\n",
        "    \n",
        "    # Strip the model of its pruning parameters-\n",
        "    winning_ticket_model_stripped = sparsity.strip_pruning(winning_ticket_model)\n",
        "    \n",
        "    # For each layer, for each weight which is 0, leave it, as is. And for weights which survive the pruning,\n",
        "    # reinitialize it to the value, the model received BEFORE it was trained and pruned-\n",
        "    for orig_wts, pruned_wts in zip(orig_model_stripped.trainable_weights,\n",
        "                                    winning_ticket_model_stripped.trainable_weights):\n",
        "        pruned_wts.assign(tf.where(tf.equal(pruned_wts, 0), pruned_wts, orig_wts))\n",
        "    \n",
        "    \n",
        "    # Save the weights (with pruning parameters) extracted to a file-\n",
        "    # winning_ticket_model.save_weights(\"LeNet_MNIST_Winning_Ticket.h5\", overwrite=True)\n",
        "    \n",
        "    \n",
        "    \n",
        "    "
      ],
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 1\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.2171, Accuracy: 93.5783, Test Loss: 0.1154, Test Accuracy: 96.630005\n",
            "Total number of trainable parameters = 266610\n",
            "\n",
            "Epoch 2, Loss: 0.0842, Accuracy: 97.3450, Test Loss: 0.0859, Test Accuracy: 97.389999\n",
            "Total number of trainable parameters = 266610\n",
            "\n",
            "Epoch 3, Loss: 0.0578, Accuracy: 98.1433, Test Loss: 0.0773, Test Accuracy: 97.509995\n",
            "Total number of trainable parameters = 266610\n",
            "\n",
            "Epoch 4, Loss: 0.0429, Accuracy: 98.6250, Test Loss: 0.0767, Test Accuracy: 97.909996\n",
            "Total number of trainable parameters = 266610\n",
            "\n",
            "Epoch 5, Loss: 0.0332, Accuracy: 98.9083, Test Loss: 0.0911, Test Accuracy: 97.500000\n",
            "Total number of trainable parameters = 266610\n",
            "\n",
            "Epoch 6, Loss: 0.0260, Accuracy: 99.1667, Test Loss: 0.0813, Test Accuracy: 98.000000\n",
            "Total number of trainable parameters = 266610\n",
            "\n",
            "Epoch 7, Loss: 0.0239, Accuracy: 99.1933, Test Loss: 0.0781, Test Accuracy: 97.799995\n",
            "Total number of trainable parameters = 266610\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 165us/sample - loss: 0.0231 - accuracy: 0.9921 - val_loss: 0.0954 - val_accuracy: 0.9784\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 145us/sample - loss: 0.0118 - accuracy: 0.9961 - val_loss: 0.0912 - val_accuracy: 0.9791\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 142us/sample - loss: 0.0104 - accuracy: 0.9964 - val_loss: 0.0829 - val_accuracy: 0.9817\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 147us/sample - loss: 0.0101 - accuracy: 0.9966 - val_loss: 0.0954 - val_accuracy: 0.9793\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 8s 138us/sample - loss: 0.0092 - accuracy: 0.9967 - val_loss: 0.1210 - val_accuracy: 0.9753\n",
            "Epoch 6/100\n",
            "60000/60000 [==============================] - 10s 161us/sample - loss: 0.0098 - accuracy: 0.9971 - val_loss: 0.0953 - val_accuracy: 0.9815\n",
            "\n",
            "Round = 1, total number of trainable parameters = 213370\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 2\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1557, Accuracy: 95.1983, Test Loss: 0.0989, Test Accuracy: 96.919998\n",
            "Total number of trainable parameters = 213370\n",
            "\n",
            "Epoch 2, Loss: 0.0613, Accuracy: 98.0183, Test Loss: 0.0746, Test Accuracy: 97.669998\n",
            "Total number of trainable parameters = 213370\n",
            "\n",
            "Epoch 3, Loss: 0.0408, Accuracy: 98.6850, Test Loss: 0.0779, Test Accuracy: 97.689995\n",
            "Total number of trainable parameters = 213370\n",
            "\n",
            "Epoch 4, Loss: 0.0276, Accuracy: 99.1167, Test Loss: 0.0759, Test Accuracy: 97.879997\n",
            "Total number of trainable parameters = 213370\n",
            "\n",
            "Epoch 5, Loss: 0.0206, Accuracy: 99.3683, Test Loss: 0.0822, Test Accuracy: 97.680000\n",
            "Total number of trainable parameters = 213370\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 166us/sample - loss: 0.0191 - accuracy: 0.9937 - val_loss: 0.0685 - val_accuracy: 0.9829\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 143us/sample - loss: 0.0097 - accuracy: 0.9968 - val_loss: 0.0765 - val_accuracy: 0.9808\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 149us/sample - loss: 0.0095 - accuracy: 0.9969 - val_loss: 0.0897 - val_accuracy: 0.9805\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 150us/sample - loss: 0.0089 - accuracy: 0.9970 - val_loss: 0.0936 - val_accuracy: 0.9794\n",
            "\n",
            "Round = 2, total number of trainable parameters = 170778\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 3\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1382, Accuracy: 95.8383, Test Loss: 0.0798, Test Accuracy: 97.479996\n",
            "Total number of trainable parameters = 170771\n",
            "\n",
            "Epoch 2, Loss: 0.0519, Accuracy: 98.3800, Test Loss: 0.0675, Test Accuracy: 97.899994\n",
            "Total number of trainable parameters = 170771\n",
            "\n",
            "Epoch 3, Loss: 0.0327, Accuracy: 98.9367, Test Loss: 0.0720, Test Accuracy: 97.820000\n",
            "Total number of trainable parameters = 170771\n",
            "\n",
            "Epoch 4, Loss: 0.0218, Accuracy: 99.2900, Test Loss: 0.0774, Test Accuracy: 97.709999\n",
            "Total number of trainable parameters = 170771\n",
            "\n",
            "Epoch 5, Loss: 0.0171, Accuracy: 99.4683, Test Loss: 0.0678, Test Accuracy: 98.119995\n",
            "Total number of trainable parameters = 170771\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 172us/sample - loss: 0.0151 - accuracy: 0.9950 - val_loss: 0.0849 - val_accuracy: 0.9802\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 148us/sample - loss: 0.0082 - accuracy: 0.9976 - val_loss: 0.0774 - val_accuracy: 0.9803\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 149us/sample - loss: 0.0070 - accuracy: 0.9979 - val_loss: 0.0898 - val_accuracy: 0.9814\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 8s 139us/sample - loss: 0.0053 - accuracy: 0.9983 - val_loss: 0.0960 - val_accuracy: 0.9800\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 9s 148us/sample - loss: 0.0053 - accuracy: 0.9982 - val_loss: 0.0910 - val_accuracy: 0.9816\n",
            "\n",
            "Round = 3, total number of trainable parameters = 136700\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 4\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1260, Accuracy: 96.1617, Test Loss: 0.0786, Test Accuracy: 97.589996\n",
            "Total number of trainable parameters = 136693\n",
            "\n",
            "Epoch 2, Loss: 0.0433, Accuracy: 98.6183, Test Loss: 0.0678, Test Accuracy: 97.949997\n",
            "Total number of trainable parameters = 136693\n",
            "\n",
            "Epoch 3, Loss: 0.0262, Accuracy: 99.1600, Test Loss: 0.0629, Test Accuracy: 98.019997\n",
            "Total number of trainable parameters = 136693\n",
            "\n",
            "Epoch 4, Loss: 0.0166, Accuracy: 99.4767, Test Loss: 0.0728, Test Accuracy: 98.150002\n",
            "Total number of trainable parameters = 136693\n",
            "\n",
            "Epoch 5, Loss: 0.0133, Accuracy: 99.5783, Test Loss: 0.0690, Test Accuracy: 98.119995\n",
            "Total number of trainable parameters = 136693\n",
            "\n",
            "Epoch 6, Loss: 0.0106, Accuracy: 99.6433, Test Loss: 0.0751, Test Accuracy: 98.159996\n",
            "Total number of trainable parameters = 136693\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 11s 176us/sample - loss: 0.0099 - accuracy: 0.9965 - val_loss: 0.0915 - val_accuracy: 0.9791\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 146us/sample - loss: 0.0052 - accuracy: 0.9986 - val_loss: 0.0755 - val_accuracy: 0.9821\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 8s 139us/sample - loss: 0.0039 - accuracy: 0.9989 - val_loss: 0.0845 - val_accuracy: 0.9819\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 8s 136us/sample - loss: 0.0037 - accuracy: 0.9989 - val_loss: 0.0919 - val_accuracy: 0.9820\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 8s 129us/sample - loss: 0.0061 - accuracy: 0.9981 - val_loss: 0.1046 - val_accuracy: 0.9808\n",
            "\n",
            "Round = 4, total number of trainable parameters = 109437\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 5\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1171, Accuracy: 96.4467, Test Loss: 0.0682, Test Accuracy: 97.820000\n",
            "Total number of trainable parameters = 109434\n",
            "\n",
            "Epoch 2, Loss: 0.0367, Accuracy: 98.8217, Test Loss: 0.0584, Test Accuracy: 98.269997\n",
            "Total number of trainable parameters = 109434\n",
            "\n",
            "Epoch 3, Loss: 0.0207, Accuracy: 99.3867, Test Loss: 0.0585, Test Accuracy: 98.180000\n",
            "Total number of trainable parameters = 109434\n",
            "\n",
            "Epoch 4, Loss: 0.0139, Accuracy: 99.5717, Test Loss: 0.0640, Test Accuracy: 98.209999\n",
            "Total number of trainable parameters = 109434\n",
            "\n",
            "Epoch 5, Loss: 0.0109, Accuracy: 99.6650, Test Loss: 0.0838, Test Accuracy: 97.949997\n",
            "Total number of trainable parameters = 109434\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 167us/sample - loss: 0.0086 - accuracy: 0.9971 - val_loss: 0.0778 - val_accuracy: 0.9795\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 10s 165us/sample - loss: 0.0057 - accuracy: 0.9982 - val_loss: 0.0726 - val_accuracy: 0.9820\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 8s 138us/sample - loss: 0.0049 - accuracy: 0.9986 - val_loss: 0.0778 - val_accuracy: 0.9824\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 158us/sample - loss: 0.0032 - accuracy: 0.9991 - val_loss: 0.0850 - val_accuracy: 0.9811\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 8s 128us/sample - loss: 0.0040 - accuracy: 0.9987 - val_loss: 0.1005 - val_accuracy: 0.9799\n",
            "\n",
            "Round = 5, total number of trainable parameters = 87631\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 6\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1062, Accuracy: 96.8550, Test Loss: 0.0659, Test Accuracy: 97.939995\n",
            "Total number of trainable parameters = 87628\n",
            "\n",
            "Epoch 2, Loss: 0.0311, Accuracy: 99.0467, Test Loss: 0.0581, Test Accuracy: 98.189995\n",
            "Total number of trainable parameters = 87628\n",
            "\n",
            "Epoch 3, Loss: 0.0171, Accuracy: 99.4950, Test Loss: 0.0628, Test Accuracy: 98.199997\n",
            "Total number of trainable parameters = 87628\n",
            "\n",
            "Epoch 4, Loss: 0.0104, Accuracy: 99.7083, Test Loss: 0.0639, Test Accuracy: 98.309998\n",
            "Total number of trainable parameters = 87628\n",
            "\n",
            "Epoch 5, Loss: 0.0073, Accuracy: 99.7983, Test Loss: 0.0757, Test Accuracy: 98.199997\n",
            "Total number of trainable parameters = 87628\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 168us/sample - loss: 0.0071 - accuracy: 0.9978 - val_loss: 0.0790 - val_accuracy: 0.9821\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 145us/sample - loss: 0.0047 - accuracy: 0.9991 - val_loss: 0.0775 - val_accuracy: 0.9816\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 144us/sample - loss: 0.0033 - accuracy: 0.9992 - val_loss: 0.0906 - val_accuracy: 0.9810\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 155us/sample - loss: 0.0035 - accuracy: 0.9990 - val_loss: 0.0872 - val_accuracy: 0.9817\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 9s 158us/sample - loss: 0.0016 - accuracy: 0.9996 - val_loss: 0.0954 - val_accuracy: 0.9816\n",
            "\n",
            "Round = 6, total number of trainable parameters = 70187\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 7\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1038, Accuracy: 96.9950, Test Loss: 0.0616, Test Accuracy: 98.049995\n",
            "Total number of trainable parameters = 70184\n",
            "\n",
            "Epoch 2, Loss: 0.0290, Accuracy: 99.1250, Test Loss: 0.0585, Test Accuracy: 98.199997\n",
            "Total number of trainable parameters = 70184\n",
            "\n",
            "Epoch 3, Loss: 0.0157, Accuracy: 99.5600, Test Loss: 0.0636, Test Accuracy: 98.159996\n",
            "Total number of trainable parameters = 70184\n",
            "\n",
            "Epoch 4, Loss: 0.0100, Accuracy: 99.7133, Test Loss: 0.0619, Test Accuracy: 98.320000\n",
            "Total number of trainable parameters = 70184\n",
            "\n",
            "Epoch 5, Loss: 0.0060, Accuracy: 99.8583, Test Loss: 0.0631, Test Accuracy: 98.290001\n",
            "Total number of trainable parameters = 70184\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 171us/sample - loss: 0.0058 - accuracy: 0.9984 - val_loss: 0.0745 - val_accuracy: 0.9822\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 10s 170us/sample - loss: 0.0050 - accuracy: 0.9988 - val_loss: 0.0689 - val_accuracy: 0.9831\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 8s 137us/sample - loss: 0.0020 - accuracy: 0.9997 - val_loss: 0.0740 - val_accuracy: 0.9822\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 8s 141us/sample - loss: 0.0015 - accuracy: 0.9998 - val_loss: 0.0848 - val_accuracy: 0.9832\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 9s 143us/sample - loss: 0.0036 - accuracy: 0.9989 - val_loss: 0.0829 - val_accuracy: 0.9833\n",
            "\n",
            "Round = 7, total number of trainable parameters = 56236\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 8\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1074, Accuracy: 96.9367, Test Loss: 0.0606, Test Accuracy: 98.099998\n",
            "Total number of trainable parameters = 56235\n",
            "\n",
            "Epoch 2, Loss: 0.0293, Accuracy: 99.1533, Test Loss: 0.0533, Test Accuracy: 98.419998\n",
            "Total number of trainable parameters = 56235\n",
            "\n",
            "Epoch 3, Loss: 0.0165, Accuracy: 99.5567, Test Loss: 0.0603, Test Accuracy: 98.209999\n",
            "Total number of trainable parameters = 56235\n",
            "\n",
            "Epoch 4, Loss: 0.0095, Accuracy: 99.7717, Test Loss: 0.0613, Test Accuracy: 98.320000\n",
            "Total number of trainable parameters = 56235\n",
            "\n",
            "Epoch 5, Loss: 0.0058, Accuracy: 99.8717, Test Loss: 0.0658, Test Accuracy: 98.269997\n",
            "Total number of trainable parameters = 56235\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 165us/sample - loss: 0.0055 - accuracy: 0.9986 - val_loss: 0.0711 - val_accuracy: 0.9832\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 157us/sample - loss: 0.0055 - accuracy: 0.9987 - val_loss: 0.0664 - val_accuracy: 0.9839\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 8s 140us/sample - loss: 0.0022 - accuracy: 0.9998 - val_loss: 0.0713 - val_accuracy: 0.9836\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 8s 140us/sample - loss: 0.0014 - accuracy: 0.9998 - val_loss: 0.0775 - val_accuracy: 0.9838\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 10s 161us/sample - loss: 0.0032 - accuracy: 0.9991 - val_loss: 0.0934 - val_accuracy: 0.9806\n",
            "\n",
            "Round = 8, total number of trainable parameters = 45068\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 9\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1153, Accuracy: 96.8483, Test Loss: 0.0656, Test Accuracy: 98.019997\n",
            "Total number of trainable parameters = 45063\n",
            "\n",
            "Epoch 2, Loss: 0.0333, Accuracy: 99.0333, Test Loss: 0.0612, Test Accuracy: 98.019997\n",
            "Total number of trainable parameters = 45063\n",
            "\n",
            "Epoch 3, Loss: 0.0188, Accuracy: 99.5050, Test Loss: 0.0580, Test Accuracy: 98.309998\n",
            "Total number of trainable parameters = 45063\n",
            "\n",
            "Epoch 4, Loss: 0.0114, Accuracy: 99.7267, Test Loss: 0.0593, Test Accuracy: 98.290001\n",
            "Total number of trainable parameters = 45063\n",
            "\n",
            "Epoch 5, Loss: 0.0072, Accuracy: 99.8217, Test Loss: 0.0648, Test Accuracy: 98.209999\n",
            "Total number of trainable parameters = 45063\n",
            "\n",
            "Epoch 6, Loss: 0.0045, Accuracy: 99.9117, Test Loss: 0.0700, Test Accuracy: 98.169998\n",
            "Total number of trainable parameters = 45063\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 169us/sample - loss: 0.0042 - accuracy: 0.9989 - val_loss: 0.0744 - val_accuracy: 0.9824\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 153us/sample - loss: 0.0045 - accuracy: 0.9991 - val_loss: 0.0713 - val_accuracy: 0.9826\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 152us/sample - loss: 0.0017 - accuracy: 0.9999 - val_loss: 0.0767 - val_accuracy: 0.9826\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 147us/sample - loss: 0.0016 - accuracy: 0.9997 - val_loss: 0.0864 - val_accuracy: 0.9819\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 9s 155us/sample - loss: 0.0021 - accuracy: 0.9995 - val_loss: 0.0817 - val_accuracy: 0.9819\n",
            "\n",
            "Round = 9, total number of trainable parameters = 36132\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 10\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1246, Accuracy: 96.7517, Test Loss: 0.0675, Test Accuracy: 97.879997\n",
            "Total number of trainable parameters = 36126\n",
            "\n",
            "Epoch 2, Loss: 0.0369, Accuracy: 98.9500, Test Loss: 0.0627, Test Accuracy: 97.939995\n",
            "Total number of trainable parameters = 36126\n",
            "\n",
            "Epoch 3, Loss: 0.0218, Accuracy: 99.3933, Test Loss: 0.0590, Test Accuracy: 98.129997\n",
            "Total number of trainable parameters = 36126\n",
            "\n",
            "Epoch 4, Loss: 0.0135, Accuracy: 99.6483, Test Loss: 0.0583, Test Accuracy: 98.269997\n",
            "Total number of trainable parameters = 36126\n",
            "\n",
            "Epoch 5, Loss: 0.0087, Accuracy: 99.7950, Test Loss: 0.0635, Test Accuracy: 98.269997\n",
            "Total number of trainable parameters = 36126\n",
            "\n",
            "Epoch 6, Loss: 0.0056, Accuracy: 99.8733, Test Loss: 0.0667, Test Accuracy: 98.239998\n",
            "Total number of trainable parameters = 36126\n",
            "\n",
            "Epoch 7, Loss: 0.0039, Accuracy: 99.9367, Test Loss: 0.0737, Test Accuracy: 98.209999\n",
            "Total number of trainable parameters = 36126\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 9s 146us/sample - loss: 0.0036 - accuracy: 0.9993 - val_loss: 0.0817 - val_accuracy: 0.9815\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 156us/sample - loss: 0.0058 - accuracy: 0.9987 - val_loss: 0.0709 - val_accuracy: 0.9821\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 149us/sample - loss: 0.0020 - accuracy: 0.9998 - val_loss: 0.0740 - val_accuracy: 0.9826\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 152us/sample - loss: 0.0012 - accuracy: 1.0000 - val_loss: 0.0821 - val_accuracy: 0.9835\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 9s 149us/sample - loss: 0.0014 - accuracy: 0.9997 - val_loss: 0.1009 - val_accuracy: 0.9793\n",
            "\n",
            "Round = 10, total number of trainable parameters = 28977\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 11\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1358, Accuracy: 96.3867, Test Loss: 0.0713, Test Accuracy: 97.839996\n",
            "Total number of trainable parameters = 28970\n",
            "\n",
            "Epoch 2, Loss: 0.0432, Accuracy: 98.7233, Test Loss: 0.0605, Test Accuracy: 98.150002\n",
            "Total number of trainable parameters = 28970\n",
            "\n",
            "Epoch 3, Loss: 0.0267, Accuracy: 99.2100, Test Loss: 0.0573, Test Accuracy: 98.309998\n",
            "Total number of trainable parameters = 28970\n",
            "\n",
            "Epoch 4, Loss: 0.0178, Accuracy: 99.5450, Test Loss: 0.0622, Test Accuracy: 98.269997\n",
            "Total number of trainable parameters = 28970\n",
            "\n",
            "Epoch 5, Loss: 0.0123, Accuracy: 99.7017, Test Loss: 0.0616, Test Accuracy: 98.189995\n",
            "Total number of trainable parameters = 28970\n",
            "\n",
            "Epoch 6, Loss: 0.0083, Accuracy: 99.8383, Test Loss: 0.0660, Test Accuracy: 98.309998\n",
            "Total number of trainable parameters = 28970\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 167us/sample - loss: 0.0067 - accuracy: 0.9983 - val_loss: 0.0761 - val_accuracy: 0.9811\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 144us/sample - loss: 0.0086 - accuracy: 0.9979 - val_loss: 0.0713 - val_accuracy: 0.9818\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 157us/sample - loss: 0.0041 - accuracy: 0.9996 - val_loss: 0.0720 - val_accuracy: 0.9825\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 157us/sample - loss: 0.0028 - accuracy: 0.9997 - val_loss: 0.0785 - val_accuracy: 0.9820\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 10s 159us/sample - loss: 0.0021 - accuracy: 0.9998 - val_loss: 0.0889 - val_accuracy: 0.9826\n",
            "\n",
            "Round = 11, total number of trainable parameters = 23256\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 12\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1566, Accuracy: 95.9000, Test Loss: 0.0807, Test Accuracy: 97.639999\n",
            "Total number of trainable parameters = 23240\n",
            "\n",
            "Epoch 2, Loss: 0.0511, Accuracy: 98.4833, Test Loss: 0.0647, Test Accuracy: 98.110001\n",
            "Total number of trainable parameters = 23240\n",
            "\n",
            "Epoch 3, Loss: 0.0324, Accuracy: 99.0850, Test Loss: 0.0617, Test Accuracy: 98.169998\n",
            "Total number of trainable parameters = 23240\n",
            "\n",
            "Epoch 4, Loss: 0.0221, Accuracy: 99.3867, Test Loss: 0.0636, Test Accuracy: 98.089996\n",
            "Total number of trainable parameters = 23240\n",
            "\n",
            "Epoch 5, Loss: 0.0165, Accuracy: 99.5783, Test Loss: 0.0640, Test Accuracy: 98.209999\n",
            "Total number of trainable parameters = 23240\n",
            "\n",
            "Epoch 6, Loss: 0.0115, Accuracy: 99.7267, Test Loss: 0.0663, Test Accuracy: 98.309998\n",
            "Total number of trainable parameters = 23240\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 9s 151us/sample - loss: 0.0094 - accuracy: 0.9978 - val_loss: 0.0707 - val_accuracy: 0.9815\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 8s 139us/sample - loss: 0.0123 - accuracy: 0.9973 - val_loss: 0.0725 - val_accuracy: 0.9822\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 8s 139us/sample - loss: 0.0069 - accuracy: 0.9987 - val_loss: 0.0722 - val_accuracy: 0.9826\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 147us/sample - loss: 0.0049 - accuracy: 0.9993 - val_loss: 0.0800 - val_accuracy: 0.9808\n",
            "\n",
            "Round = 12, total number of trainable parameters = 18667\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 13\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1680, Accuracy: 95.6200, Test Loss: 0.0902, Test Accuracy: 97.379997\n",
            "Total number of trainable parameters = 18652\n",
            "\n",
            "Epoch 2, Loss: 0.0590, Accuracy: 98.2233, Test Loss: 0.0745, Test Accuracy: 97.719994\n",
            "Total number of trainable parameters = 18652\n",
            "\n",
            "Epoch 3, Loss: 0.0399, Accuracy: 98.7933, Test Loss: 0.0690, Test Accuracy: 98.059998\n",
            "Total number of trainable parameters = 18652\n",
            "\n",
            "Epoch 4, Loss: 0.0289, Accuracy: 99.1683, Test Loss: 0.0650, Test Accuracy: 98.139999\n",
            "Total number of trainable parameters = 18652\n",
            "\n",
            "Epoch 5, Loss: 0.0221, Accuracy: 99.4250, Test Loss: 0.0688, Test Accuracy: 98.139999\n",
            "Total number of trainable parameters = 18652\n",
            "\n",
            "Epoch 6, Loss: 0.0170, Accuracy: 99.5433, Test Loss: 0.0668, Test Accuracy: 98.239998\n",
            "Total number of trainable parameters = 18652\n",
            "\n",
            "Epoch 7, Loss: 0.0131, Accuracy: 99.6683, Test Loss: 0.0727, Test Accuracy: 98.239998\n",
            "Total number of trainable parameters = 18652\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 9s 150us/sample - loss: 0.0108 - accuracy: 0.9971 - val_loss: 0.0767 - val_accuracy: 0.9815\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 143us/sample - loss: 0.0270 - accuracy: 0.9925 - val_loss: 0.0736 - val_accuracy: 0.9799\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 8s 140us/sample - loss: 0.0128 - accuracy: 0.9971 - val_loss: 0.0758 - val_accuracy: 0.9794\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 8s 141us/sample - loss: 0.0096 - accuracy: 0.9980 - val_loss: 0.0789 - val_accuracy: 0.9801\n",
            "Epoch 5/100\n",
            "60000/60000 [==============================] - 8s 140us/sample - loss: 0.0078 - accuracy: 0.9985 - val_loss: 0.0809 - val_accuracy: 0.9803\n",
            "\n",
            "Round = 13, total number of trainable parameters = 14993\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 14\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.1928, Accuracy: 94.7400, Test Loss: 0.1005, Test Accuracy: 97.129997\n",
            "Total number of trainable parameters = 14967\n",
            "\n",
            "Epoch 2, Loss: 0.0730, Accuracy: 97.8733, Test Loss: 0.0812, Test Accuracy: 97.689995\n",
            "Total number of trainable parameters = 14967\n",
            "\n",
            "Epoch 3, Loss: 0.0524, Accuracy: 98.3967, Test Loss: 0.0769, Test Accuracy: 97.680000\n",
            "Total number of trainable parameters = 14967\n",
            "\n",
            "Epoch 4, Loss: 0.0407, Accuracy: 98.7833, Test Loss: 0.0719, Test Accuracy: 97.839996\n",
            "Total number of trainable parameters = 14967\n",
            "\n",
            "Epoch 5, Loss: 0.0326, Accuracy: 99.0533, Test Loss: 0.0723, Test Accuracy: 98.040001\n",
            "Total number of trainable parameters = 14967\n",
            "\n",
            "Epoch 6, Loss: 0.0263, Accuracy: 99.2417, Test Loss: 0.0704, Test Accuracy: 98.159996\n",
            "Total number of trainable parameters = 14967\n",
            "\n",
            "Epoch 7, Loss: 0.0221, Accuracy: 99.3900, Test Loss: 0.0740, Test Accuracy: 98.079994\n",
            "Total number of trainable parameters = 14967\n",
            "\n",
            "Epoch 8, Loss: 0.0184, Accuracy: 99.4800, Test Loss: 0.0747, Test Accuracy: 98.049995\n",
            "Total number of trainable parameters = 14967\n",
            "\n",
            "Epoch 9, Loss: 0.0160, Accuracy: 99.5750, Test Loss: 0.0769, Test Accuracy: 98.019997\n",
            "Total number of trainable parameters = 14967\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 173us/sample - loss: 0.0136 - accuracy: 0.9962 - val_loss: 0.0806 - val_accuracy: 0.9803\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 152us/sample - loss: 0.1899 - accuracy: 0.9670 - val_loss: 0.1361 - val_accuracy: 0.9730\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 10s 160us/sample - loss: 0.0726 - accuracy: 0.9872 - val_loss: 0.1016 - val_accuracy: 0.9770\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 150us/sample - loss: 0.0485 - accuracy: 0.9903 - val_loss: 0.0897 - val_accuracy: 0.9778\n",
            "\n",
            "Round = 14, total number of trainable parameters = 12039\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 15\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.3521, Accuracy: 92.0717, Test Loss: 0.1627, Test Accuracy: 96.559998\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 2, Loss: 0.1168, Accuracy: 97.5083, Test Loss: 0.1147, Test Accuracy: 97.290001\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 3, Loss: 0.0798, Accuracy: 98.1850, Test Loss: 0.0959, Test Accuracy: 97.570000\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 4, Loss: 0.0605, Accuracy: 98.5833, Test Loss: 0.0862, Test Accuracy: 97.689995\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 5, Loss: 0.0485, Accuracy: 98.8117, Test Loss: 0.0833, Test Accuracy: 97.750000\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 6, Loss: 0.0400, Accuracy: 98.9733, Test Loss: 0.0813, Test Accuracy: 97.739998\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 7, Loss: 0.0339, Accuracy: 99.1150, Test Loss: 0.0807, Test Accuracy: 97.820000\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 8, Loss: 0.0290, Accuracy: 99.2733, Test Loss: 0.0803, Test Accuracy: 97.839996\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 9, Loss: 0.0254, Accuracy: 99.3517, Test Loss: 0.0810, Test Accuracy: 97.839996\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 10, Loss: 0.0221, Accuracy: 99.4450, Test Loss: 0.0803, Test Accuracy: 97.919998\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 11, Loss: 0.0195, Accuracy: 99.5350, Test Loss: 0.0850, Test Accuracy: 97.889999\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 12, Loss: 0.0172, Accuracy: 99.5850, Test Loss: 0.0879, Test Accuracy: 97.809998\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "Epoch 13, Loss: 0.0154, Accuracy: 99.6383, Test Loss: 0.0904, Test Accuracy: 97.739998\n",
            "Total number of trainable parameters = 12008\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 169us/sample - loss: 0.0144 - accuracy: 0.9966 - val_loss: 0.0895 - val_accuracy: 0.9789\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 152us/sample - loss: 0.0910 - accuracy: 0.9797 - val_loss: 0.1114 - val_accuracy: 0.9727\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 154us/sample - loss: 0.0472 - accuracy: 0.9894 - val_loss: 0.0998 - val_accuracy: 0.9753\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 10s 160us/sample - loss: 0.0359 - accuracy: 0.9916 - val_loss: 0.0944 - val_accuracy: 0.9762\n",
            "\n",
            "Round = 15, total number of trainable parameters = 9665\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 16\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.4586, Accuracy: 91.5050, Test Loss: 0.2094, Test Accuracy: 96.360001\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 2, Loss: 0.1571, Accuracy: 96.9167, Test Loss: 0.1393, Test Accuracy: 96.959999\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 3, Loss: 0.1059, Accuracy: 97.6517, Test Loss: 0.1117, Test Accuracy: 97.360001\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 4, Loss: 0.0816, Accuracy: 98.1000, Test Loss: 0.1005, Test Accuracy: 97.439995\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 5, Loss: 0.0660, Accuracy: 98.3983, Test Loss: 0.0939, Test Accuracy: 97.549995\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 6, Loss: 0.0564, Accuracy: 98.5650, Test Loss: 0.0887, Test Accuracy: 97.610001\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 7, Loss: 0.0489, Accuracy: 98.7650, Test Loss: 0.0878, Test Accuracy: 97.639999\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 8, Loss: 0.0432, Accuracy: 98.8783, Test Loss: 0.0869, Test Accuracy: 97.669998\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 9, Loss: 0.0390, Accuracy: 98.9767, Test Loss: 0.0859, Test Accuracy: 97.709999\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 10, Loss: 0.0358, Accuracy: 99.0750, Test Loss: 0.0857, Test Accuracy: 97.649994\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 11, Loss: 0.0326, Accuracy: 99.1383, Test Loss: 0.0871, Test Accuracy: 97.709999\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 12, Loss: 0.0300, Accuracy: 99.2350, Test Loss: 0.0897, Test Accuracy: 97.669998\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "Epoch 13, Loss: 0.0278, Accuracy: 99.2733, Test Loss: 0.0902, Test Accuracy: 97.629997\n",
            "Total number of trainable parameters = 9644\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 163us/sample - loss: 0.0260 - accuracy: 0.9932 - val_loss: 0.0940 - val_accuracy: 0.9764\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 147us/sample - loss: 0.1617 - accuracy: 0.9721 - val_loss: 0.1480 - val_accuracy: 0.9671\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 144us/sample - loss: 0.0781 - accuracy: 0.9838 - val_loss: 0.1245 - val_accuracy: 0.9707\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 8s 141us/sample - loss: 0.0588 - accuracy: 0.9864 - val_loss: 0.1157 - val_accuracy: 0.9720\n",
            "\n",
            "Round = 16, total number of trainable parameters = 7774\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 17\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.6075, Accuracy: 90.6667, Test Loss: 0.3185, Test Accuracy: 95.810005\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 2, Loss: 0.2408, Accuracy: 96.5900, Test Loss: 0.2112, Test Accuracy: 96.620003\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 3, Loss: 0.1642, Accuracy: 97.2817, Test Loss: 0.1680, Test Accuracy: 96.860001\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 4, Loss: 0.1257, Accuracy: 97.6700, Test Loss: 0.1426, Test Accuracy: 96.970001\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 5, Loss: 0.1015, Accuracy: 97.9450, Test Loss: 0.1290, Test Accuracy: 97.040001\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 6, Loss: 0.0859, Accuracy: 98.1483, Test Loss: 0.1199, Test Accuracy: 97.159996\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 7, Loss: 0.0750, Accuracy: 98.2833, Test Loss: 0.1157, Test Accuracy: 97.129997\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 8, Loss: 0.0668, Accuracy: 98.4083, Test Loss: 0.1093, Test Accuracy: 97.259995\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 9, Loss: 0.0606, Accuracy: 98.5367, Test Loss: 0.1071, Test Accuracy: 97.320000\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 10, Loss: 0.0560, Accuracy: 98.6633, Test Loss: 0.1061, Test Accuracy: 97.330002\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 11, Loss: 0.0517, Accuracy: 98.6617, Test Loss: 0.1052, Test Accuracy: 97.339996\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 12, Loss: 0.0486, Accuracy: 98.7617, Test Loss: 0.1063, Test Accuracy: 97.239998\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 13, Loss: 0.0457, Accuracy: 98.8567, Test Loss: 0.1060, Test Accuracy: 97.290001\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "Epoch 14, Loss: 0.0431, Accuracy: 98.8983, Test Loss: 0.1070, Test Accuracy: 97.339996\n",
            "Total number of trainable parameters = 7753\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 175us/sample - loss: 0.0415 - accuracy: 0.9888 - val_loss: 0.1066 - val_accuracy: 0.9731\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 153us/sample - loss: 0.0583 - accuracy: 0.9832 - val_loss: 0.1176 - val_accuracy: 0.9694\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 10s 167us/sample - loss: 0.0490 - accuracy: 0.9860 - val_loss: 0.1140 - val_accuracy: 0.9718\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 153us/sample - loss: 0.0462 - accuracy: 0.9870 - val_loss: 0.1154 - val_accuracy: 0.9723\n",
            "\n",
            "Round = 17, total number of trainable parameters = 6255\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 18\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.6126, Accuracy: 90.0600, Test Loss: 0.3190, Test Accuracy: 95.570000\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "Epoch 2, Loss: 0.2449, Accuracy: 96.3783, Test Loss: 0.2154, Test Accuracy: 96.300003\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "Epoch 3, Loss: 0.1699, Accuracy: 97.0367, Test Loss: 0.1711, Test Accuracy: 96.699997\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "Epoch 4, Loss: 0.1311, Accuracy: 97.4650, Test Loss: 0.1462, Test Accuracy: 96.959999\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "Epoch 5, Loss: 0.1075, Accuracy: 97.7700, Test Loss: 0.1320, Test Accuracy: 96.959999\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "Epoch 6, Loss: 0.0921, Accuracy: 97.9450, Test Loss: 0.1254, Test Accuracy: 97.010002\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "Epoch 7, Loss: 0.0809, Accuracy: 98.1067, Test Loss: 0.1193, Test Accuracy: 97.180000\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "Epoch 8, Loss: 0.0731, Accuracy: 98.2117, Test Loss: 0.1184, Test Accuracy: 97.099998\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "Epoch 9, Loss: 0.0672, Accuracy: 98.3267, Test Loss: 0.1134, Test Accuracy: 97.139999\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "Epoch 10, Loss: 0.0626, Accuracy: 98.3950, Test Loss: 0.1138, Test Accuracy: 97.089996\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "Epoch 11, Loss: 0.0584, Accuracy: 98.4867, Test Loss: 0.1141, Test Accuracy: 97.129997\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "Epoch 12, Loss: 0.0553, Accuracy: 98.5283, Test Loss: 0.1137, Test Accuracy: 97.099998\n",
            "Total number of trainable parameters = 6241\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 172us/sample - loss: 0.0525 - accuracy: 0.9860 - val_loss: 0.1166 - val_accuracy: 0.9721\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 153us/sample - loss: 0.0868 - accuracy: 0.9746 - val_loss: 0.1277 - val_accuracy: 0.9679\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 153us/sample - loss: 0.0718 - accuracy: 0.9795 - val_loss: 0.1255 - val_accuracy: 0.9694\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 144us/sample - loss: 0.0678 - accuracy: 0.9809 - val_loss: 0.1226 - val_accuracy: 0.9699\n",
            "\n",
            "Round = 18, total number of trainable parameters = 5039\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 19\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 0.6644, Accuracy: 89.0350, Test Loss: 0.3512, Test Accuracy: 95.099998\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 2, Loss: 0.2718, Accuracy: 95.6517, Test Loss: 0.2380, Test Accuracy: 95.880005\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 3, Loss: 0.1927, Accuracy: 96.3383, Test Loss: 0.1910, Test Accuracy: 96.270004\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 4, Loss: 0.1528, Accuracy: 96.8033, Test Loss: 0.1665, Test Accuracy: 96.459999\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 5, Loss: 0.1291, Accuracy: 97.0583, Test Loss: 0.1509, Test Accuracy: 96.680000\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 6, Loss: 0.1131, Accuracy: 97.2817, Test Loss: 0.1425, Test Accuracy: 96.879997\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 7, Loss: 0.1021, Accuracy: 97.4717, Test Loss: 0.1353, Test Accuracy: 96.770004\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 8, Loss: 0.0938, Accuracy: 97.6200, Test Loss: 0.1308, Test Accuracy: 96.919998\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 9, Loss: 0.0877, Accuracy: 97.6617, Test Loss: 0.1286, Test Accuracy: 96.959999\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 10, Loss: 0.0827, Accuracy: 97.8117, Test Loss: 0.1279, Test Accuracy: 96.900002\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 11, Loss: 0.0790, Accuracy: 97.8533, Test Loss: 0.1257, Test Accuracy: 96.839996\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 12, Loss: 0.0758, Accuracy: 97.9517, Test Loss: 0.1266, Test Accuracy: 96.839996\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 13, Loss: 0.0730, Accuracy: 98.0033, Test Loss: 0.1251, Test Accuracy: 96.900002\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 14, Loss: 0.0706, Accuracy: 98.0383, Test Loss: 0.1253, Test Accuracy: 96.970001\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 15, Loss: 0.0686, Accuracy: 98.1017, Test Loss: 0.1268, Test Accuracy: 96.849998\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 16, Loss: 0.0665, Accuracy: 98.1750, Test Loss: 0.1250, Test Accuracy: 96.900002\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 17, Loss: 0.0646, Accuracy: 98.2083, Test Loss: 0.1255, Test Accuracy: 96.879997\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 18, Loss: 0.0632, Accuracy: 98.2250, Test Loss: 0.1268, Test Accuracy: 96.849998\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "Epoch 19, Loss: 0.0619, Accuracy: 98.2233, Test Loss: 0.1265, Test Accuracy: 96.820000\n",
            "Total number of trainable parameters = 5005\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 164us/sample - loss: 0.0605 - accuracy: 0.9825 - val_loss: 0.1281 - val_accuracy: 0.9684\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 155us/sample - loss: 1.6861 - accuracy: 0.7427 - val_loss: 1.3944 - val_accuracy: 0.7468\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 150us/sample - loss: 1.0969 - accuracy: 0.7591 - val_loss: 0.9775 - val_accuracy: 0.7503\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 146us/sample - loss: 0.7997 - accuracy: 0.7614 - val_loss: 0.7644 - val_accuracy: 0.7530\n",
            "\n",
            "Round = 19, total number of trainable parameters = 4046\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 20\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 1.0003, Accuracy: 70.4633, Test Loss: 0.6908, Test Accuracy: 75.720001\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 2, Loss: 0.6241, Accuracy: 76.0350, Test Loss: 0.5854, Test Accuracy: 76.120003\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 3, Loss: 0.5527, Accuracy: 76.5367, Test Loss: 0.5435, Test Accuracy: 76.270004\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 4, Loss: 0.5189, Accuracy: 76.7900, Test Loss: 0.5238, Test Accuracy: 76.430000\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 5, Loss: 0.4991, Accuracy: 76.9967, Test Loss: 0.5109, Test Accuracy: 76.450005\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 6, Loss: 0.4863, Accuracy: 77.1000, Test Loss: 0.5030, Test Accuracy: 76.610001\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 7, Loss: 0.4773, Accuracy: 77.2117, Test Loss: 0.4980, Test Accuracy: 76.580002\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 8, Loss: 0.4708, Accuracy: 77.2617, Test Loss: 0.4942, Test Accuracy: 76.639999\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 9, Loss: 0.4657, Accuracy: 77.3583, Test Loss: 0.4935, Test Accuracy: 76.670006\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 10, Loss: 0.4619, Accuracy: 77.3583, Test Loss: 0.4901, Test Accuracy: 76.650002\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 11, Loss: 0.4588, Accuracy: 77.4083, Test Loss: 0.4900, Test Accuracy: 76.639999\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 12, Loss: 0.4559, Accuracy: 77.4800, Test Loss: 0.4891, Test Accuracy: 76.670006\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 13, Loss: 0.4538, Accuracy: 77.5583, Test Loss: 0.4887, Test Accuracy: 76.570000\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 14, Loss: 0.4519, Accuracy: 77.5700, Test Loss: 0.4893, Test Accuracy: 76.660004\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 15, Loss: 0.4502, Accuracy: 77.6000, Test Loss: 0.4870, Test Accuracy: 76.810005\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 16, Loss: 0.4485, Accuracy: 77.6200, Test Loss: 0.4899, Test Accuracy: 76.730003\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 17, Loss: 0.4472, Accuracy: 77.6367, Test Loss: 0.4873, Test Accuracy: 76.810005\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "Epoch 18, Loss: 0.4459, Accuracy: 77.6833, Test Loss: 0.4886, Test Accuracy: 76.760002\n",
            "Total number of trainable parameters = 4004\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 9s 150us/sample - loss: 0.4452 - accuracy: 0.7766 - val_loss: 0.4875 - val_accuracy: 0.7675\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 145us/sample - loss: 1.6805 - accuracy: 0.6337 - val_loss: 1.2527 - val_accuracy: 0.6464\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 143us/sample - loss: 0.9563 - accuracy: 0.6554 - val_loss: 0.8277 - val_accuracy: 0.6490\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 8s 141us/sample - loss: 0.7283 - accuracy: 0.6576 - val_loss: 0.7391 - val_accuracy: 0.6486\n",
            "\n",
            "Round = 20, total number of trainable parameters = 3239\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 21\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 1.2100, Accuracy: 61.2200, Test Loss: 0.9323, Test Accuracy: 65.729996\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 2, Loss: 0.8568, Accuracy: 66.4350, Test Loss: 0.8203, Test Accuracy: 66.529999\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 3, Loss: 0.7861, Accuracy: 66.8783, Test Loss: 0.7815, Test Accuracy: 66.649994\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 4, Loss: 0.7537, Accuracy: 67.1717, Test Loss: 0.7586, Test Accuracy: 66.809998\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 5, Loss: 0.7354, Accuracy: 67.2933, Test Loss: 0.7472, Test Accuracy: 66.809998\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 6, Loss: 0.7231, Accuracy: 67.4100, Test Loss: 0.7400, Test Accuracy: 66.869995\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 7, Loss: 0.7145, Accuracy: 67.5017, Test Loss: 0.7334, Test Accuracy: 67.019997\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 8, Loss: 0.7078, Accuracy: 67.6050, Test Loss: 0.7310, Test Accuracy: 67.049995\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 9, Loss: 0.7031, Accuracy: 67.6317, Test Loss: 0.7254, Test Accuracy: 67.079994\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 10, Loss: 0.6990, Accuracy: 67.7517, Test Loss: 0.7226, Test Accuracy: 67.139999\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 11, Loss: 0.6956, Accuracy: 67.7583, Test Loss: 0.7223, Test Accuracy: 67.059998\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 12, Loss: 0.6930, Accuracy: 67.8200, Test Loss: 0.7224, Test Accuracy: 67.180000\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 13, Loss: 0.6912, Accuracy: 67.8267, Test Loss: 0.7210, Test Accuracy: 67.159996\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 14, Loss: 0.6890, Accuracy: 67.8667, Test Loss: 0.7218, Test Accuracy: 67.029999\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 15, Loss: 0.6875, Accuracy: 67.9183, Test Loss: 0.7222, Test Accuracy: 67.139999\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 16, Loss: 0.6860, Accuracy: 67.8983, Test Loss: 0.7208, Test Accuracy: 67.119995\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 17, Loss: 0.6846, Accuracy: 67.8917, Test Loss: 0.7226, Test Accuracy: 67.199997\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 18, Loss: 0.6835, Accuracy: 67.9400, Test Loss: 0.7216, Test Accuracy: 67.070000\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "Epoch 19, Loss: 0.6822, Accuracy: 67.9400, Test Loss: 0.7236, Test Accuracy: 67.169998\n",
            "Total number of trainable parameters = 3219\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 9s 150us/sample - loss: 0.6814 - accuracy: 0.6798 - val_loss: 0.7258 - val_accuracy: 0.6718\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 145us/sample - loss: 1.0566 - accuracy: 0.5778 - val_loss: 1.0083 - val_accuracy: 0.5729\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 147us/sample - loss: 0.9344 - accuracy: 0.5813 - val_loss: 0.9618 - val_accuracy: 0.5741\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 157us/sample - loss: 0.9155 - accuracy: 0.5822 - val_loss: 0.9585 - val_accuracy: 0.5753\n",
            "\n",
            "Round = 21, total number of trainable parameters = 2604\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 22\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 1.3702, Accuracy: 52.9033, Test Loss: 1.1246, Test Accuracy: 56.199997\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 2, Loss: 1.0595, Accuracy: 56.9017, Test Loss: 1.0355, Test Accuracy: 56.550003\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 3, Loss: 1.0001, Accuracy: 57.3083, Test Loss: 1.0019, Test Accuracy: 57.110001\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 4, Loss: 0.9740, Accuracy: 57.5650, Test Loss: 0.9864, Test Accuracy: 57.209999\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 5, Loss: 0.9585, Accuracy: 57.7050, Test Loss: 0.9778, Test Accuracy: 57.349998\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 6, Loss: 0.9481, Accuracy: 57.8450, Test Loss: 0.9712, Test Accuracy: 57.389999\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 7, Loss: 0.9408, Accuracy: 57.9717, Test Loss: 0.9668, Test Accuracy: 57.349998\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 8, Loss: 0.9352, Accuracy: 58.0150, Test Loss: 0.9639, Test Accuracy: 57.470001\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 9, Loss: 0.9310, Accuracy: 58.0767, Test Loss: 0.9596, Test Accuracy: 57.639999\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 10, Loss: 0.9276, Accuracy: 58.1600, Test Loss: 0.9606, Test Accuracy: 57.470001\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 11, Loss: 0.9255, Accuracy: 58.1917, Test Loss: 0.9589, Test Accuracy: 57.560001\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 12, Loss: 0.9229, Accuracy: 58.2300, Test Loss: 0.9584, Test Accuracy: 57.660000\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 13, Loss: 0.9211, Accuracy: 58.2733, Test Loss: 0.9601, Test Accuracy: 57.580002\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 14, Loss: 0.9200, Accuracy: 58.2683, Test Loss: 0.9580, Test Accuracy: 57.599998\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 15, Loss: 0.9187, Accuracy: 58.2650, Test Loss: 0.9602, Test Accuracy: 57.500000\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 16, Loss: 0.9176, Accuracy: 58.2683, Test Loss: 0.9595, Test Accuracy: 57.500000\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "Epoch 17, Loss: 0.9167, Accuracy: 58.2917, Test Loss: 0.9584, Test Accuracy: 57.570000\n",
            "Total number of trainable parameters = 2585\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 169us/sample - loss: 0.9162 - accuracy: 0.5833 - val_loss: 0.9581 - val_accuracy: 0.5773\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 10s 161us/sample - loss: 1.0506 - accuracy: 0.5621 - val_loss: 1.0099 - val_accuracy: 0.5677\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 153us/sample - loss: 0.9780 - accuracy: 0.5715 - val_loss: 0.9982 - val_accuracy: 0.5689\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 10s 168us/sample - loss: 0.9677 - accuracy: 0.5730 - val_loss: 0.9936 - val_accuracy: 0.5694\n",
            "\n",
            "Round = 22, total number of trainable parameters = 2094\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 23\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 1.4215, Accuracy: 51.6550, Test Loss: 1.1618, Test Accuracy: 55.500000\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 2, Loss: 1.1016, Accuracy: 56.2050, Test Loss: 1.0676, Test Accuracy: 56.279999\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 3, Loss: 1.0390, Accuracy: 56.6050, Test Loss: 1.0330, Test Accuracy: 56.529999\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 4, Loss: 1.0110, Accuracy: 56.8983, Test Loss: 1.0158, Test Accuracy: 56.739998\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 5, Loss: 0.9955, Accuracy: 57.0517, Test Loss: 1.0060, Test Accuracy: 56.889999\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 6, Loss: 0.9862, Accuracy: 57.1783, Test Loss: 1.0019, Test Accuracy: 56.840004\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 7, Loss: 0.9797, Accuracy: 57.2817, Test Loss: 0.9964, Test Accuracy: 56.950005\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 8, Loss: 0.9751, Accuracy: 57.3333, Test Loss: 0.9962, Test Accuracy: 56.900002\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 9, Loss: 0.9717, Accuracy: 57.3633, Test Loss: 0.9938, Test Accuracy: 57.090004\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 10, Loss: 0.9688, Accuracy: 57.4333, Test Loss: 0.9958, Test Accuracy: 56.779999\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 11, Loss: 0.9663, Accuracy: 57.4667, Test Loss: 0.9913, Test Accuracy: 56.989998\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 12, Loss: 0.9645, Accuracy: 57.4733, Test Loss: 0.9900, Test Accuracy: 57.040001\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 13, Loss: 0.9630, Accuracy: 57.4667, Test Loss: 0.9912, Test Accuracy: 56.970001\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 14, Loss: 0.9617, Accuracy: 57.5283, Test Loss: 0.9891, Test Accuracy: 57.010002\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 15, Loss: 0.9608, Accuracy: 57.5267, Test Loss: 0.9884, Test Accuracy: 56.989998\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 16, Loss: 0.9594, Accuracy: 57.5167, Test Loss: 0.9884, Test Accuracy: 57.090004\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 17, Loss: 0.9586, Accuracy: 57.5583, Test Loss: 0.9894, Test Accuracy: 57.059998\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 18, Loss: 0.9579, Accuracy: 57.5517, Test Loss: 0.9886, Test Accuracy: 57.099998\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 19, Loss: 0.9571, Accuracy: 57.5767, Test Loss: 0.9878, Test Accuracy: 57.040001\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 20, Loss: 0.9565, Accuracy: 57.5817, Test Loss: 0.9892, Test Accuracy: 57.000000\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 21, Loss: 0.9560, Accuracy: 57.5667, Test Loss: 0.9879, Test Accuracy: 57.090004\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "Epoch 22, Loss: 0.9556, Accuracy: 57.6217, Test Loss: 0.9903, Test Accuracy: 57.029999\n",
            "Total number of trainable parameters = 2075\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 10s 170us/sample - loss: 0.9553 - accuracy: 0.5765 - val_loss: 0.9899 - val_accuracy: 0.5703\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 9s 153us/sample - loss: 0.9778 - accuracy: 0.5722 - val_loss: 0.9959 - val_accuracy: 0.5690\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 9s 154us/sample - loss: 0.9671 - accuracy: 0.5737 - val_loss: 0.9933 - val_accuracy: 0.5693\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 9s 148us/sample - loss: 0.9654 - accuracy: 0.5741 - val_loss: 0.9936 - val_accuracy: 0.5693\n",
            "\n",
            "Round = 23, total number of trainable parameters = 1682\n",
            "\n",
            "\n",
            "\n",
            "\n",
            "Iterative pruning round: 24\n",
            "\n",
            "\n",
            "Epoch 1, Loss: 1.4370, Accuracy: 51.4567, Test Loss: 1.1710, Test Accuracy: 55.500000\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 2, Loss: 1.1091, Accuracy: 55.9717, Test Loss: 1.0716, Test Accuracy: 56.019997\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 3, Loss: 1.0471, Accuracy: 56.4233, Test Loss: 1.0357, Test Accuracy: 56.320000\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 4, Loss: 1.0195, Accuracy: 56.6983, Test Loss: 1.0196, Test Accuracy: 56.699997\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 5, Loss: 1.0043, Accuracy: 56.8567, Test Loss: 1.0108, Test Accuracy: 56.680000\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 6, Loss: 0.9950, Accuracy: 56.9967, Test Loss: 1.0054, Test Accuracy: 56.800003\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 7, Loss: 0.9891, Accuracy: 57.0800, Test Loss: 1.0056, Test Accuracy: 56.650002\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 8, Loss: 0.9846, Accuracy: 57.1100, Test Loss: 0.9997, Test Accuracy: 56.739998\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 9, Loss: 0.9811, Accuracy: 57.1683, Test Loss: 1.0006, Test Accuracy: 56.809998\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 10, Loss: 0.9787, Accuracy: 57.2033, Test Loss: 0.9962, Test Accuracy: 56.889999\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 11, Loss: 0.9765, Accuracy: 57.2217, Test Loss: 0.9947, Test Accuracy: 56.870003\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 12, Loss: 0.9745, Accuracy: 57.2450, Test Loss: 0.9956, Test Accuracy: 56.840004\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 13, Loss: 0.9730, Accuracy: 57.2900, Test Loss: 0.9945, Test Accuracy: 56.849998\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 14, Loss: 0.9718, Accuracy: 57.2950, Test Loss: 0.9948, Test Accuracy: 56.900002\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 15, Loss: 0.9710, Accuracy: 57.2667, Test Loss: 0.9928, Test Accuracy: 56.910004\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 16, Loss: 0.9698, Accuracy: 57.3333, Test Loss: 0.9938, Test Accuracy: 56.900002\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 17, Loss: 0.9692, Accuracy: 57.2917, Test Loss: 0.9939, Test Accuracy: 56.889999\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "Epoch 18, Loss: 0.9682, Accuracy: 57.3067, Test Loss: 0.9933, Test Accuracy: 56.889999\n",
            "Total number of trainable parameters = 1672\n",
            "\n",
            "\n",
            "'EarlyStopping' called!\n",
            "\n",
            "Train on 60000 samples, validate on 10000 samples\n",
            "Epoch 1/100\n",
            "60000/60000 [==============================] - 9s 150us/sample - loss: 0.9679 - accuracy: 0.5730 - val_loss: 0.9953 - val_accuracy: 0.5682\n",
            "Epoch 2/100\n",
            "60000/60000 [==============================] - 8s 138us/sample - loss: 3.3220 - accuracy: 0.3433 - val_loss: 2.5002 - val_accuracy: 0.4392\n",
            "Epoch 3/100\n",
            "60000/60000 [==============================] - 7s 118us/sample - loss: 1.9695 - accuracy: 0.4475 - val_loss: 1.6133 - val_accuracy: 0.4492\n",
            "Epoch 4/100\n",
            "60000/60000 [==============================] - 7s 120us/sample - loss: 1.4210 - accuracy: 0.4531 - val_loss: 1.3406 - val_accuracy: 0.4509\n",
            "\n",
            "Round = 24, total number of trainable parameters = 1358\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YJjPHrOIIwvr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zPQEVJqFIwvw",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QNO6ll_TY3LH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Save the weights (with pruning parameters) AFTER completion of 'n'\n",
        "# iterative pruning rounds-\n",
        "winning_ticket_model.save_weights(\"LeNet_MNIST_Winning_Ticket.h5\", overwrite=True)  "
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OoBQIIxnY-us",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_jSDmA0OIwvz",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "outputId": "d7259b92-474f-442f-b44b-547ff07943ce"
      },
      "source": [
        "print(\"\\nIterative-pruning for LeNet-300-100-10 using OPTIMIZED Lottery Ticket Hypothesis is now complete.\\n\")"
      ],
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Iterative-pruning for LeNet-300-100-10 using OPTIMIZED Lottery Ticket Hypothesis is now complete.\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5uVHHjx1Iwv3",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AMG44joXIwv6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "szWuur40Iwv8",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import os\n",
        "import pickle"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PYvmSBUKIwwA",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "9f73d7d8-ecc2-47ee-dd78-24f043adbd43"
      },
      "source": [
        "os.getcwd()"
      ],
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'/content'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 56
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bGuJSIdUIwwF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Save/Pickle Python 3 dictionary-\n",
        "\n",
        "# with open(\"/home/majumdar/The_Lottery_Ticket_Hypothesis-Resources/LeNet_MNIST_history_main_Experiment_3.pkl\", \"wb\") as f:\n",
        "with open(\"/content/LeNet_MNIST_history_main_Optimized_Code.pkl\", \"wb\") as f:\n",
        "    pickle.dump(history_main, f)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "s7wp12JvIwwY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gCGGm0O8Iwwg",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Load pickled object as Python 3 dictionary-\n",
        "\n",
        "# with open(\"/home/majumdar/The_Lottery_Ticket_Hypothesis-Resources/LeNet_MNIST_history_main_Experiment_3.pkl\", \"rb\") as f:\n",
        "with open(\"/home/arjun/Desktop/Codes/Lottery_Hypothesis-Resources/Latest_Works/LTH_Experiments/Experiment_number_1/LeNet_MNIST_history_main.pkl\", \"rb\") as f:\n",
        "    history_main = pickle.load(f)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Xq33uSBAIwwj",
        "colab_type": "code",
        "colab": {},
        "outputId": "7472284e-5122-47e4-c623-95ecf1fce9a8"
      },
      "source": [
        "# Sanity check-\n",
        "history_main[10]['percentage_wts_pruned']"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "89.13094032481902"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "v8Mtz8HbIwwn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XUEVTK9fIwwp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PDUp5QxQIwwt",
        "colab_type": "text"
      },
      "source": [
        "### Visualizations for _Iterative Pruning Rounds_:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XsSb0x3nIwwx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Python 3 dict for training and testing accuracy visualization-\n",
        "# key is 'percentage_wts_pruned' and value is accuracy\n",
        "plot_accuracy = {}\n",
        "plot_test_accuracy = {}"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y6HnYZGSIww2",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# populate 'plot_accuracy'-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['accuracy'])\n",
        "    plot_accuracy[history_main[k]['percentage_wts_pruned']] = history_main[k]['accuracy'][epoch_length - 1]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5-8MID9UIww6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# populate 'plot_test_accuracy'-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['accuracy'])\n",
        "    plot_test_accuracy[history_main[k]['percentage_wts_pruned']] = history_main[k]['val_accuracy'][epoch_length - 1]\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gMsoYWHiIww8",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 504
        },
        "outputId": "68cea66f-a081-4be6-93d2-f4a16bfcd4cc"
      },
      "source": [
        "# Visualization of training and testing accuracy VS percentage of weights\n",
        "# pruned-\n",
        "fig=plt.figure(figsize=(9, 7), dpi= 80, facecolor='w', edgecolor='k')\n",
        "\n",
        "plt.plot(list(plot_accuracy.keys()), list(plot_accuracy.values()), label = 'training_accuracy')\n",
        "plt.plot(list(plot_test_accuracy.keys()), list(plot_test_accuracy.values()), label = 'testing_accuracy')\n",
        "\n",
        "plt.title(\"LeNet-300-100 : Percentage of weights pruned VS Accuracy\")\n",
        "plt.xlabel(\"percentage of weights pruned\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend(loc = 'best')\n",
        "plt.show()"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmkAAAHnCAYAAADjI8rnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nUkoppVQNpEWaUkoppVQN9L9q4jRo0rVM3QAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 720x560 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QCgt1vgCIwxC",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GnUXKct5IwxE",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uiEhc3UWIwxG",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Python 3 dict for training and testing loss visualization-\n",
        "plot_loss = {}\n",
        "plot_test_loss = {}"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "q-O2YSk9IwxK",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# populate 'plot_loss'-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['loss'])\n",
        "    plot_loss[history_main[k]['percentage_wts_pruned']] = history_main[k]['loss'][epoch_length - 1]\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jqzeE71uIwxN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# populate 'plot_test_loss'-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['loss'])\n",
        "    plot_test_loss[history_main[k]['percentage_wts_pruned']] = history_main[k]['val_loss'][epoch_length - 1]\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3JDgWgdkIwxT",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 504
        },
        "outputId": "40986089-c260-4f9f-9387-8fc5ddc34083"
      },
      "source": [
        "# Visualization of training and testing loss VS percentage of remaining weights-\n",
        "# pruned-\n",
        "fig=plt.figure(figsize=(9, 7), dpi= 80, facecolor='w', edgecolor='k')\n",
        "\n",
        "plt.plot(list(plot_loss.keys()), list(plot_loss.values()), label = 'training_loss')\n",
        "plt.plot(list(plot_test_loss.keys()), list(plot_test_loss.values()), label = 'testing_loss')\n",
        "\n",
        "plt.title(\"LeNet-300-100: Percentage of weights pruned VS Loss\")\n",
        "plt.xlabel(\"percentage of weights pruned\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend(loc = 'best')\n",
        "plt.show()"
      ],
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmUAAAHnCAYAAAD5HyppAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nMeYEOJQxxhhjjDkBDmWMMcYYY07g/wFFCr58NNkY5gAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 720x560 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-1iw2sO0IwxZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EibDzbZ3Iwxd",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Plot number of epochs (Early Stopping) VS percentage of weights pruned-\n",
        "\n",
        "# Python 3 dict to hold number of epochs vs % of weights pruned-\n",
        "plot_num_epochs = {}\n",
        "plot_num_epochs_test = {}\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CNQq8CDQIwxh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# populate 'plot_num_epochs'-\n",
        "for k in history_main.keys():\n",
        "    num_epochs = len(history_main[k]['accuracy'])\n",
        "    plot_num_epochs[history_main[k]['percentage_wts_pruned']] = num_epochs\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4qSXBieXIwxj",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "'''\n",
        "# populate 'plot_num_epochs_test'-\n",
        "for k in history_main.keys():\n",
        "    num_epochs = len(history_main[k]['val_accuracy'])\n",
        "    plot_num_epochs_test[history[k]['percentage_wts_pruned']] = num_epochs\n",
        "'''"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7KQtpC1PIwxm",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 504
        },
        "outputId": "24637e11-aa2d-4516-8e0f-5b929cfcde1f"
      },
      "source": [
        "# Visualize percentage of weights remaining VS number of epochs (Early Stopping)\n",
        "fig=plt.figure(figsize=(9, 7), dpi= 80, facecolor='w', edgecolor='k')\n",
        "\n",
        "plt.plot(list(plot_num_epochs.keys()), list(plot_num_epochs.values()), label = 'training set')\n",
        "# plt.plot(list(plot_num_epochs_test.keys()), list(plot_num_epochs_test.values()), label = 'testing set')\n",
        "\n",
        "plt.title(\"LeNet-300-100 : Percentage of weights pruned VS number of epochs (Early Stopping) until convergence\")\n",
        "plt.xlabel(\"percentage of weights pruned\")\n",
        "plt.ylabel(\"number of epochs\")\n",
        "plt.legend(loc = 'best')\n",
        "plt.show()\n"
      ],
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAssAAAHnCAYAAABOjhALAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nzLTv7mIAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 720x560 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m4KD3wfKIwxq",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "urIg7ZdyIwxs",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nwG4PWfOIwxv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Try to visualize the STARTING training and testing accuracy VS percentage\n",
        "# of remaining weights-\n",
        "\n",
        "# Python 3 dict for visualization-\n",
        "plot_starting_accuracy = {}\n",
        "plot_starting_test_accuracy = {}\n",
        "\n",
        "# Key:    percentage of weights pruned\n",
        "# Value:  starting accuracy\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "U3y0v0d7Iwxy",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# populate 'plot_starting_accuracy'-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['loss'])\n",
        "    plot_starting_accuracy[history_main[k]['percentage_wts_pruned']] = history_main[k]['accuracy'][0]\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4mjXwpQoIwx1",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# populate 'plot_starting_test_accuracy'-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['loss'])\n",
        "    plot_starting_test_accuracy[history_main[k]['percentage_wts_pruned']] = history_main[k]['val_accuracy'][0]\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "PYWTBNqEIwx4",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 504
        },
        "outputId": "9f910a2f-a4a3-4f14-ba29-254b615b3779"
      },
      "source": [
        "# Visualize starting accuracy VS percentage of weights pruned-\n",
        "fig=plt.figure(figsize=(9, 7), dpi= 80, facecolor='w', edgecolor='k')\n",
        "\n",
        "plt.plot(list(plot_starting_accuracy.keys()), list(plot_starting_accuracy.values()), label = 'training starting accuracy')\n",
        "plt.plot(list(plot_starting_test_accuracy.keys()), list(plot_starting_test_accuracy.values()), label = 'testing starting accuracy')\n",
        "\n",
        "plt.title(\"LeNet-300-100 : Percentage of weights pruned VS Starting Accuracy\")\n",
        "plt.xlabel(\"percentage of weights pruned\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend(loc = 'best')\n",
        "plt.show()\n"
      ],
      "execution_count": 72,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmkAAAHnCAYAAADjI8rnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n0oQQQggh3ND/AYbpiCCtKdUjAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 720x560 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gPAJgdYtIwx7",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IVWgmpRVIwx_",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xPzBBKOfIwyB",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Try to visualize the STARTING training and testing loss VS percentage\n",
        "# of remaining weights-\n",
        "\n",
        "# Python 3 dict for visualization-\n",
        "plot_starting_loss = {}\n",
        "plot_starting_test_loss = {}\n",
        "\n",
        "# Key:\t\tpercentage of weights pruned\n",
        "# Value:\tstarting loss\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5byhZzG3IwyD",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Populate 'plot_starting_loss' Python 3 dict-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['loss'])\n",
        "    plot_starting_loss[history_main[k]['percentage_wts_pruned']] = history_main[k]['loss'][0]\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IiDrYpLLIwyH",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Populate 'plot_starting_test_loss' Python 3 dict-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['loss'])\n",
        "    plot_starting_test_loss[history_main[k]['percentage_wts_pruned']] = history_main[k]['val_loss'][0]\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QHLvyf7jIwyJ",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 504
        },
        "outputId": "e2faf006-b84a-4513-8b11-5824b184ebf2"
      },
      "source": [
        "# Visualize Starting training & testing loss VS percentage of weights pruned-\n",
        "fig=plt.figure(figsize=(9, 7), dpi= 80, facecolor='w', edgecolor='k')\n",
        "\n",
        "plt.plot(list(plot_starting_loss.keys()), list(plot_starting_loss.values()), label = 'training starting loss')\n",
        "plt.plot(list(plot_starting_test_loss.keys()), list(plot_starting_test_loss.values()), label = 'testing starting loss')\n",
        "\n",
        "plt.title(\"LeNet-300-100 : Percentage of weights pruned VS Starting Loss\")\n",
        "plt.xlabel(\"percentage of weights pruned\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend(loc = 'best')\n",
        "plt.show()"
      ],
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAmUAAAHnCAYAAAD5HyppAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZURERERB4P8DA5lONcMtJdYAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 720x560 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oH3l90VeIwyN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3PJg42_AIwyO",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1iLRkhTAIwyR",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Try to visualize: percentage of weights remaining VS BEST accuracy within each iterative\n",
        "# pruning round\n",
        "\n",
        "# Python 3 dict for visualization-\n",
        "plot_best_accuracy = {}\n",
        "plot_best_test_accuracy = {}\n",
        "\n",
        "# Key:\t\tpercentage of weights pruned\n",
        "# Value:\tbest accuracy"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nA7vcv-VIwyV",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Populate 'plot_best_accuracy' Python 3 dict-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['accuracy'])\n",
        "    plot_best_accuracy[history_main[k]['percentage_wts_pruned']] = np.amax(history_main[k]['accuracy'])\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-l8bfoFKIwyY",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Populate 'plot_best_test_accuracy' Python 3 dict-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['accuracy'])\n",
        "    plot_best_test_accuracy[history_main[k]['percentage_wts_pruned']] = np.amax(history_main[k]['val_accuracy'])\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "x_261VKVIwyb",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 625
        },
        "outputId": "0e1787a6-6a5f-4fff-e299-a8de042d4272"
      },
      "source": [
        "# Visualize best accuracy VS percentage of weights pruned-\n",
        "fig=plt.figure(figsize=(10, 9), dpi= 80, facecolor='w', edgecolor='k')\n",
        "\n",
        "plt.plot(list(plot_best_accuracy.keys()), list(plot_best_accuracy.values()), label = 'training best accuracy')\n",
        "plt.plot(list(plot_best_test_accuracy.keys()), list(plot_best_test_accuracy.values()), label = 'testing best accuracy')\n",
        "\n",
        "plt.title(\"LeNet-300-100 : Percentage of weights pruned VS Best Accuracy\")\n",
        "plt.xlabel(\"percentage of weights pruned\")\n",
        "plt.ylabel(\"Accuracy\")\n",
        "plt.legend(loc = 'best')\n",
        "plt.show()\n"
      ],
      "execution_count": 80,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAqcAAAJgCAYAAAC3CMoIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 800x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lokvHu4cIwyf",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wuLKwqnsIwyh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S-PtGTJhIwyk",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Try to visualize: percentage of weights remaining VS LOWEST loss within each iterative\n",
        "# pruning round\n",
        "\n",
        "# Python 3 dict for visualization-\n",
        "plot_best_loss = {}\n",
        "plot_best_test_loss = {}\n",
        "\n",
        "# Key:\t\tpercentage of weights pruned\n",
        "# Value:\tbest loss"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NLvdnzSiIwyn",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Populate 'plot_best_loss' Python 3 dict-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['loss'])\n",
        "    plot_best_loss[history_main[k]['percentage_wts_pruned']] = np.amin(history_main[k]['loss'])\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-ehvoECQIwyp",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# Populate 'plot_best_test_loss' Python 3 dict-\n",
        "for k in history_main.keys():\n",
        "    epoch_length = len(history_main[k]['loss'])\n",
        "    plot_best_test_loss[history_main[k]['percentage_wts_pruned']] = np.amin(history_main[k]['val_loss'])\n"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xMifjVXcIwys",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 625
        },
        "outputId": "da9e4b49-c283-4c04-e760-575ccb9d987f"
      },
      "source": [
        "# Visualize best loss VS percentage of weights pruned-\n",
        "fig=plt.figure(figsize=(10, 9), dpi= 80, facecolor='w', edgecolor='k')\n",
        "\n",
        "plt.plot(list(plot_best_loss.keys()), list(plot_best_loss.values()), label = 'training best loss')\n",
        "plt.plot(list(plot_best_test_loss.keys()), list(plot_best_test_loss.values()), label = 'testing best loss')\n",
        "\n",
        "plt.title(\"LeNet-300-100: Percentage of weights pruned VS Best Loss\")\n",
        "plt.xlabel(\"percentage of weights pruned\")\n",
        "plt.ylabel(\"Loss\")\n",
        "plt.legend(loc = 'best')\n",
        "plt.show()\n"
      ],
      "execution_count": 85,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAqMAAAJgCAYAAAC+42pyAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 800x720 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lWif1_xEIwyv",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_nrbWi5nIwyx",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "5Cs9NL6iIwy0",
        "colab_type": "text"
      },
      "source": [
        "### Find the iterative pruning round for which training and testing accuracy is highest:"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LjIOlLu3Iwy2",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "outputId": "ee4d0c39-2801-4894-b743-3432951b930f"
      },
      "source": [
        "# Find highest training and testing accuracy for all iterative pruning rounds-\n",
        "best_accuracy = 0\n",
        "iterative_round = 0\n",
        "\n",
        "\n",
        "for k in history_main.keys():\n",
        "\tepoch_length = len(history_main[k]['accuracy'])\n",
        "\tif history_main[k]['accuracy'][epoch_length - 1] > best_accuracy:\n",
        "\t\tbest_accuracy = history_main[k]['accuracy'][epoch_length - 1]\n",
        "\t\titerative_round = k\n",
        "\n",
        "\n",
        "print(\"\\nIterative round = {0} has highest training accuracy = {1:.4f}% at percentage of weights pruned = {2:.4f}%\\n\".format(\n",
        "    iterative_round, best_accuracy, history_main[iterative_round]['percentage_wts_pruned']))\n"
      ],
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Iterative round = 10 has highest training accuracy = 99.9367% at percentage of weights pruned = 89.1313%\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dAznRl22Iwy9",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 69
        },
        "outputId": "537be942-0faf-49ce-d480-1d83d5bed28d"
      },
      "source": [
        "best_test_accuracy = 0\n",
        "iterative_round = 0\n",
        "\n",
        "\n",
        "for k in history_main.keys():\n",
        "\tepoch_length = len(history_main[k]['val_accuracy'])\n",
        "\tif history_main[k]['val_accuracy'][epoch_length - 1] > best_test_accuracy:\n",
        "\t\tbest_test_accuracy = history_main[k]['val_accuracy'][epoch_length - 1]\n",
        "\t\titerative_round = k\n",
        "\n",
        "\n",
        "print(\"\\nIterative round = {0} has highest testing accuracy = {1:.4f}% at percentage of weights pruned = {2:.4f}%\\n\".format(\n",
        "    iterative_round, best_test_accuracy, history_main[iterative_round]['percentage_wts_pruned']))\n"
      ],
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "Iterative round = 11 has highest testing accuracy = 98.3100% at percentage of weights pruned = 91.2771%\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6o29p_QoIwzX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bxg0nqSWkdtP",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OaPOyFzmkfFl",
        "colab_type": "text"
      },
      "source": [
        "### To evaluate the winning ticket, we need to see the following output:\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Yv41S3zrIwza",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 454
        },
        "outputId": "f73a91d4-4448-49da-bb15-d03d29412fb0"
      },
      "source": [
        "for k in history_main.keys():\n",
        "  epoch_length = len(history_main[k]['loss'])\n",
        "  print(\"Round = {0}: val_accuracy = {1:.4f}% & val_loss = {2:.4f} with {3:.4f}% weights pruned, while number of epochs = {4}\".format(\n",
        "      k, history_main[k]['val_accuracy'][epoch_length - 1], history_main[k]['val_loss'][epoch_length - 1], \n",
        "      history_main[k]['percentage_wts_pruned'], epoch_length\n",
        "  ))"
      ],
      "execution_count": 93,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Round = 1: val_accuracy = 97.8000% & val_loss = 0.0781 with 19.9692% weights pruned, while number of epochs = 7\n",
            "Round = 2: val_accuracy = 97.6800% & val_loss = 0.0822 with 35.9446% weights pruned, while number of epochs = 5\n",
            "Round = 3: val_accuracy = 98.1200% & val_loss = 0.0678 with 48.7266% weights pruned, while number of epochs = 5\n",
            "Round = 4: val_accuracy = 98.1600% & val_loss = 0.0751 with 58.9524% weights pruned, while number of epochs = 6\n",
            "Round = 5: val_accuracy = 97.9500% & val_loss = 0.0838 with 67.1314% weights pruned, while number of epochs = 5\n",
            "Round = 6: val_accuracy = 98.2000% & val_loss = 0.0757 with 73.6743% weights pruned, while number of epochs = 5\n",
            "Round = 7: val_accuracy = 98.2900% & val_loss = 0.0631 with 78.9070% weights pruned, while number of epochs = 5\n",
            "Round = 8: val_accuracy = 98.2700% & val_loss = 0.0658 with 83.0959% weights pruned, while number of epochs = 5\n",
            "Round = 9: val_accuracy = 98.1700% & val_loss = 0.0700 with 86.4476% weights pruned, while number of epochs = 6\n",
            "Round = 10: val_accuracy = 98.2100% & val_loss = 0.0737 with 89.1313% weights pruned, while number of epochs = 7\n",
            "Round = 11: val_accuracy = 98.3100% & val_loss = 0.0660 with 91.2771% weights pruned, while number of epochs = 6\n",
            "Round = 12: val_accuracy = 98.3100% & val_loss = 0.0663 with 92.9984% weights pruned, while number of epochs = 6\n",
            "Round = 13: val_accuracy = 98.2400% & val_loss = 0.0727 with 94.3764% weights pruned, while number of epochs = 7\n",
            "Round = 14: val_accuracy = 98.0200% & val_loss = 0.0769 with 95.4844% weights pruned, while number of epochs = 9\n",
            "Round = 15: val_accuracy = 97.7400% & val_loss = 0.0904 with 96.3749% weights pruned, while number of epochs = 13\n",
            "Round = 16: val_accuracy = 97.6300% & val_loss = 0.0902 with 97.0841% weights pruned, while number of epochs = 13\n",
            "Round = 17: val_accuracy = 97.3400% & val_loss = 0.1070 with 97.6539% weights pruned, while number of epochs = 14\n",
            "Round = 18: val_accuracy = 97.1000% & val_loss = 0.1137 with 98.1100% weights pruned, while number of epochs = 12\n",
            "Round = 19: val_accuracy = 96.8200% & val_loss = 0.1265 with 98.4824% weights pruned, while number of epochs = 19\n",
            "Round = 20: val_accuracy = 76.7600% & val_loss = 0.4886 with 98.7851% weights pruned, while number of epochs = 18\n",
            "Round = 21: val_accuracy = 67.1700% & val_loss = 0.7236 with 99.0233% weights pruned, while number of epochs = 19\n",
            "Round = 22: val_accuracy = 57.5700% & val_loss = 0.9584 with 99.2146% weights pruned, while number of epochs = 17\n",
            "Round = 23: val_accuracy = 57.0300% & val_loss = 0.9903 with 99.3691% weights pruned, while number of epochs = 22\n",
            "Round = 24: val_accuracy = 56.8900% & val_loss = 0.9933 with 99.4906% weights pruned, while number of epochs = 18\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BHxvP_56jSyb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lg5bs1YAIwzb",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 0,
      "outputs": []
    }
  ]
}