deepLearning_GAN.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "deepLearning_GAN.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/hfarruda/deeplearningtutorial/blob/master/deepLearning_GAN.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6snJPOhSX1-k"
      },
      "source": [
        "#  Generative Adversarial Networks\n",
        "This example is part of the [*Deep Learning Tutorial*](https://github.com/hfarruda/deeplearningtutorial), authored by Henrique F. de Arruda, Alexandre Benatti, César Comin, and Luciano da Fontoura Costa. This code is not suitable for other data and/or applications, which will require modifications in the structure and parameters. This code has absolutely no warranty.\n",
        "\n",
        "If you publish a paper related on this material, please cite:\n",
        "\n",
        "H. F. de Arruda, A. Benatti, C. H. Comin, L. da F. Costa, \"Learning Deep Learning (CDT-15),\" 2019.\n",
        "\n",
        "It was elaborated to create a network that can generate handwritten characters automatically.\n",
        "\n",
        "\n",
        "First of all, we import the necessary libraries. Here we opt for using Keras (using TensorFlow backend). "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "d9Y2iDK8Xtxo",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6d892661-d842-4a86-dd3c-11853add0f2a"
      },
      "source": [
        "%tensorflow_version 1.x\n",
        "import numpy as np\n",
        "import pandas as pd \n",
        "import keras\n",
        "from keras.models import Sequential, model_from_json\n",
        "from keras.utils.vis_utils import plot_model\n",
        "from keras.datasets import mnist\n",
        "from keras.layers import InputLayer, Dense, Flatten, Reshape, Input, Dropout\n",
        "from keras.layers.advanced_activations import LeakyReLU\n",
        "from keras.layers import BatchNormalization\n",
        "from keras.models import Model,Sequential\n",
        "from keras.regularizers import L1L2\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.metrics import accuracy_score\n",
        "from sklearn.preprocessing import MinMaxScaler\n",
        "import matplotlib.pyplot as plt\n",
        "import cv2 "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "TensorFlow 1.x selected.\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "Using TensorFlow backend.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YbavlsFUwLL5"
      },
      "source": [
        "If you have a GPU, you can use the following code to allocate processing into it.  Otherwise, proceed to (*)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1s4IcXtXwKms",
        "outputId": "eb3a0456-d4f5-4136-b1f5-eabab53506ea",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "import tensorflow as tf \n",
        "from keras import backend as K\n",
        "\n",
        "print(K.tensorflow_backend._get_available_gpus())\n",
        "\n",
        "number_of_cpu_cores = 8\n",
        "config = tf.ConfigProto(device_count = {'GPU': 1 , 'CPU': number_of_cpu_cores}) \n",
        "session = tf.Session(config=config) \n",
        "keras.backend.set_session(session)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['/job:localhost/replica:0/task:0/device:GPU:0']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Q34xwESqb6Yd"
      },
      "source": [
        "(*) In this example we used the MNIST database in which it is composed by grayscale images of the 10 handwritten digits. It is available at Keras library on [keras-datasets](https://keras.io/datasets/).\n",
        "\n",
        "The following command is used to load the data set."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6NVQ_qsXd9JM",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "03e1560c-798b-4b0f-8723-909b3201c55c"
      },
      "source": [
        "(train_data_raw, train_target_raw), (_, _) = mnist.load_data()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://s3.amazonaws.com/img-datasets/mnist.npz\n",
            "11493376/11490434 [==============================] - 0s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OFgXvWsDyKV-"
      },
      "source": [
        "Because this code consumes too much of processing time, here we considered only the zeros and ones."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7YYmA9nKccfz"
      },
      "source": [
        "train_data = [img for i, img in enumerate(train_data_raw) \n",
        "              if train_target_raw[i] == 0 or train_target_raw[i] == 1]\n",
        "train_data = np.array(train_data)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ddp7ztHvw5FT"
      },
      "source": [
        "In order to visualize a given figure, the following code can be executed."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "F09KdG8_fdhD",
        "outputId": "3afd9881-41a8-495e-a48f-1e976ec1e939",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 118
        }
      },
      "source": [
        "image_id = 1000\n",
        "plt.figure(figsize = (1,1))\n",
        "plt.imshow(train_data[image_id], cmap='gray')\n",
        "plt.title(\"Test image: \" + str(image_id))\n",
        "#plt.axis('off')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 72x72 with 1 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ee1Uj-yUw940"
      },
      "source": [
        "Definition of the used variables."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0FS4sGH3w-T1"
      },
      "source": [
        "input_shape = train_data.shape[1::]\n",
        "activation_output_generator = 'sigmoid'\n",
        "activation_output_discrimninator = 'sigmoid'\n",
        "input_dim = 50\n",
        "number_of_epochs = 1000\n",
        "batch_size = 100\n",
        "train_data = train_data.astype('float32') / 255"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lwNQJ042hMp5"
      },
      "source": [
        "In the following, we present the generator model."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "La0qUyS2hMWP",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "2c86a30a-24c3-4d47-e805-60451c90ca95"
      },
      "source": [
        "generator_model = Sequential()\n",
        "\n",
        "generator_model.add(Dense(units=64,input_dim = input_dim, \n",
        "                          kernel_regularizer = L1L2(1e-5, 1e-5)))\n",
        "generator_model.add(BatchNormalization())\n",
        "generator_model.add(LeakyReLU(alpha=0.3))\n",
        "\n",
        "generator_model.add(Dense(units=128, kernel_regularizer = L1L2(1e-5, 1e-5)))\n",
        "generator_model.add(BatchNormalization())\n",
        "generator_model.add(LeakyReLU(alpha=0.3))\n",
        "\n",
        "generator_model.add(Dense(units=256, kernel_regularizer = L1L2(1e-5, 1e-5)))\n",
        "generator_model.add(BatchNormalization())\n",
        "generator_model.add(LeakyReLU(alpha=0.3))\n",
        "\n",
        "generator_model.add(Dense(units = input_shape[0] * input_shape[1], \n",
        "                          activation = activation_output_generator))\n",
        "\n",
        "generator_model.add(Reshape(input_shape))\n",
        "\n",
        "generator_model.compile(loss='binary_crossentropy', optimizer=\"adam\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WARNING:tensorflow:From /tensorflow-1.15.2/python3.7/tensorflow_core/python/ops/resource_variable_ops.py:1630: calling BaseResourceVariable.__init__ (from tensorflow.python.ops.resource_variable_ops) with constraint is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "If using Keras pass *_constraint arguments to layers.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S47XpkuFhcf_"
      },
      "source": [
        "The summary of the generator model is shown by employing the following code."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Dllo6b79hj-O",
        "outputId": "0b3aaf3c-e4fe-45e4-ad06-68c02b35de3e",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "generator_model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "dense_1 (Dense)              (None, 64)                3264      \n",
            "_________________________________________________________________\n",
            "batch_normalization_1 (Batch (None, 64)                256       \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_1 (LeakyReLU)    (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 128)               8320      \n",
            "_________________________________________________________________\n",
            "batch_normalization_2 (Batch (None, 128)               512       \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_2 (LeakyReLU)    (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 256)               33024     \n",
            "_________________________________________________________________\n",
            "batch_normalization_3 (Batch (None, 256)               1024      \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_3 (LeakyReLU)    (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 784)               201488    \n",
            "_________________________________________________________________\n",
            "reshape_1 (Reshape)          (None, 28, 28)            0         \n",
            "=================================================================\n",
            "Total params: 247,888\n",
            "Trainable params: 246,992\n",
            "Non-trainable params: 896\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nKzAMwfvsxqs"
      },
      "source": [
        "The following code represents the discriminator model."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uSLEkiglh7UI",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8533f503-164e-4044-a195-aa9227e3a684"
      },
      "source": [
        "discriminator_model = Sequential()\n",
        "discriminator_model.add(InputLayer(input_shape = input_shape))\n",
        "discriminator_model.add(Flatten())\n",
        "\n",
        "discriminator_model.add(Dense(units=256,kernel_regularizer = L1L2(1e-5, 1e-5)))\n",
        "discriminator_model.add(LeakyReLU(alpha=0.3))\n",
        "discriminator_model.add(Dropout(0.2))\n",
        "\n",
        "\n",
        "discriminator_model.add(Dense(units=128,kernel_regularizer = L1L2(1e-5, 1e-5)))\n",
        "discriminator_model.add(LeakyReLU(alpha=0.3))\n",
        "discriminator_model.add(Dropout(0.2))\n",
        "\n",
        "discriminator_model.add(Dense(units=64,kernel_regularizer = L1L2(1e-5, 1e-5)))\n",
        "discriminator_model.add(LeakyReLU(alpha=0.3))\n",
        "\n",
        "discriminator_model.add(Dense(units=1, \n",
        "                              activation = activation_output_discrimninator))\n",
        "\n",
        "discriminator_model.compile(loss='binary_crossentropy', \n",
        "                            optimizer = \"adam\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "WARNING:tensorflow:From /tensorflow-1.15.2/python3.7/tensorflow_core/python/ops/nn_impl.py:183: where (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Use tf.where in 2.0, which has the same broadcast rule as np.where\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3J_e4c3SiDzI"
      },
      "source": [
        "The summary of the discriminator model is shown by using the following code."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1ouhcw-9iJcl",
        "outputId": "7f20783c-b89b-4ef5-fa14-1d0ad48676e3",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "discriminator_model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "flatten_1 (Flatten)          (None, 784)               0         \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 256)               200960    \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_4 (LeakyReLU)    (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 256)               0         \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 128)               32896     \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_5 (LeakyReLU)    (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 128)               0         \n",
            "_________________________________________________________________\n",
            "dense_7 (Dense)              (None, 64)                8256      \n",
            "_________________________________________________________________\n",
            "leaky_re_lu_6 (LeakyReLU)    (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_8 (Dense)              (None, 1)                 65        \n",
            "=================================================================\n",
            "Total params: 242,177\n",
            "Trainable params: 242,177\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "niIy42YbiSAi"
      },
      "source": [
        "The following code incorporates the complete gan model."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jokcnP7Oic6g"
      },
      "source": [
        "gan_input = Input(shape = (input_dim,))\n",
        "gan_output= discriminator_model(generator_model(gan_input))\n",
        "gan = Model(inputs = gan_input, outputs = gan_output)\n",
        "gan.compile(loss = 'binary_crossentropy', optimizer = 'adam')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Z2yCHLa2ihBW"
      },
      "source": [
        "The summary of the gan model is shown by using the following code."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6qh1aDCwihWX",
        "outputId": "8195de25-8fca-4917-c37f-c3f90deaee20",
        "colab": {
          "base_uri": "https://localhost:8080/"
        }
      },
      "source": [
        "gan.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_2 (InputLayer)         (None, 50)                0         \n",
            "_________________________________________________________________\n",
            "sequential_1 (Sequential)    (None, 28, 28)            247888    \n",
            "_________________________________________________________________\n",
            "sequential_2 (Sequential)    (None, 1)                 242177    \n",
            "=================================================================\n",
            "Total params: 490,065\n",
            "Trainable params: 489,169\n",
            "Non-trainable params: 896\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zzJedNDOivAC"
      },
      "source": [
        "Next, we train the GAN."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7-heB9oQivwj",
        "outputId": "84ad1871-d707-4e8c-fe88-9f9d9f2eda4b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 903
        }
      },
      "source": [
        "y = np.ones(batch_size)\n",
        "\n",
        "#Parameters of the noise distribution\n",
        "mu = 0\n",
        "sigma = 1\n",
        "\n",
        "#We created this array to avoid number repetitions\n",
        "train_indices = np.arange(train_data.shape[0])\n",
        "np.random.shuffle(train_indices)\n",
        "\n",
        "#Here we define the labels used to train the gan\n",
        "train_labels = np.zeros(2*batch_size,dtype = int)\n",
        "train_labels[0:batch_size] = 1#generated images\n",
        "\n",
        "for epoch in range(number_of_epochs):\n",
        "    print(\"\\rEpoch:\", epoch + 1, \"of\", number_of_epochs, end = '')\n",
        "    for _ in range(batch_size):\n",
        "        input_noise = np.random.normal(loc = mu, scale = sigma, \n",
        "                                       size = [batch_size, input_dim])\n",
        "        generated_images = generator_model.predict(input_noise)\n",
        "        np.random.shuffle(train_indices)\n",
        "        image_batch = train_data[train_indices[0:batch_size]]\n",
        "        train_images = np.concatenate((image_batch, generated_images))\n",
        "        #Training the discriminator\n",
        "        discriminator_model.trainable = True\n",
        "        discriminator_model.train_on_batch(train_images, train_labels)\n",
        "        #Training the gan\n",
        "        discriminator_model.trainable = False\n",
        "        train_noise = np.random.normal(loc = mu, scale = sigma, \n",
        "                                       size =  [batch_size, input_dim])\n",
        "        gan.train_on_batch(train_noise, y)\n",
        "\n",
        "  #In order to visualize the training progress, we employ the following code.\n",
        "    if epoch % 100 == 0:\n",
        "        n_examples = 10\n",
        "        scale_image = 1 * n_examples\n",
        "        noise= np.random.normal(loc = mu, scale = sigma, \n",
        "                                size = (n_examples, input_dim))\n",
        "        generated_images = generator_model.predict(noise)\n",
        "        n_pixels = generated_images.shape[1]\n",
        "        n_pixels_col = np.int(np.sqrt(n_pixels))\n",
        "        fig, axes = plt.subplots(1,n_examples, \n",
        "                                 figsize = (scale_image, \n",
        "                                            scale_image * n_examples))\n",
        "    for i in range(n_examples):\n",
        "        axes[i].imshow(generated_images[i], cmap = \"gray\")\n",
        "        axes[i].axis('off')\n",
        "    plt.show()\n",
        "print(\"\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\rEpoch: 1 of 1000"
          ]
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "/tensorflow-1.15.2/python3.7/keras/engine/training.py:297: UserWarning: Discrepancy between trainable weights and collected trainable weights, did you set `model.trainable` without calling `model.compile` after ?\n",
            "  'Discrepancy between trainable weights and collected trainable'\n",
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:41: DeprecationWarning: `np.int` is a deprecated alias for the builtin `int`. To silence this warning, use `int` by itself. Doing this will not modify any behavior and is safe. When replacing `np.int`, you may wish to use e.g. `np.int64` or `np.int32` to specify the precision. If you wish to review your current use, check the release note link for additional information.\n",
            "Deprecated in NumPy 1.20; for more details and guidance: https://numpy.org/devdocs/release/1.20.0-notes.html#deprecations\n"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x7200 with 10 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 101 of 1000"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x7200 with 10 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 201 of 1000"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x7200 with 10 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "metadata": {
            "tags": null
          },
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch: 301 of 1000"
          ]
        },
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x7200 with 10 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch: 401 of 1000"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x7200 with 10 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch: 501 of 1000"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x7200 with 10 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch: 601 of 1000"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x7200 with 10 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch: 701 of 1000"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x7200 with 10 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch: 801 of 1000"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x7200 with 10 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch: 901 of 1000"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAjwAAAA9CAYAAACpzLMWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO2de1BU5/nHv7vsBZY7utwChoKiIsEFjNomGp3EidTgqEkmTWqTtKNtMpVItE3SpjNtp/WXmiZqgjNNjWlHm8amOjJNNCZpnHiJokYEQQUEicpNYFkXhGV3Yc/z+4OeLbAXzi5nz176fmbO6HDOnn2/5znveZ99z/M+j4yIwGAwGAwGgxHKyP3dAAaDwWAwGAxfwxweBoPBYDAYIQ9zeBgMBoPBYIQ8zOFhMBgMBoMR8jCHh8FgMBgMRsjDHB4Gg8FgMBghj8LdTplMFtRr1olINtExTGPgM5HGUNcHMI3BANM4cogkDfEdzIb/OcznDfEtTjWyGR4Gg8FgMBghD3N4GAwGg8FghDzM4WEwGAwGgxHyMIeHwWAwGAxGyOM2aFksNBoNsrOz0dPTg5aWlpEvViiQnZ0Nk8mE69evIzw8HCqVCv39/eA4Topm+QW5XI7s7GwAQENDA7RaLdLT00FEsFqtqK+vx/DwsJ9b6ZyIiAhkZWVBoVAgLCxszD6j0Yhr164hPT0dU6ZMQUNDAwYHB/3UUvFQKpWYPXs2BgYGHPTJZDJkZ2dDr9ejtbXV3031GncaQ8GGcrkcKpUK2dnZUCqV9r/bbDY0NDTAarUiNjYWZrMZJpPJjy2dHLwdFYqRxzrHcSAiXL16NSTsGGo4GxcZPoaIXG4YidSe9KbT6chgMNDrr79u/1tSUhKdO3eO/vznPxMAmjVrFj300EMUHR0tyneOSHOtTWyNQreIiAiqqKigkydPUnh4OG3YsIFMJhMZDAaqqamhpKSkgNWYm5tLDQ0N1NXVRSaTacz24YcfEgB68803Sa/X09y5cyWzoy/tlZqaSs3NzfThhx+STCYbo0+n01FPT8+Y+zoY79PRGv1lQ19qjIyMpFmzZtH169fH3LO8xoSEBHryySdp3rx5QasRAKWkpFBTUxP19fWRwWAgvV5PN27coNzcXCntGOxIZkMhzw+ZTEZyuVzU+0SIRqkvug9wqkuSGZ6enh7s3r0bp0+fBgCsXLkSOp0OaWlpICKUlpairq4ON27cwNDQkBRN8it6vR5arRYbNmzAfffdh4iICHz11Vdobm7G2rVr0dDQgEOHDvm7mQ4oFArExMQgOjoa4eHhY/bNnj0bpaWluPfeexETE4O1a9ciJycH+/fvD9gZq/FotVqsXr0azc3NqK6uxpIlSzBnzhxMmTIFs2fPxsaNG8fok8vliI6OHjNrEAzIZDLExMQgPj4eCxYswKxZs+wag92GAJCYmIhHH30UKpUKMpkMKpUK8fHxSEhIQEREhP04hUKBtWvX4vr169Dr9UGlcTzTpk1DVlYWoqOjERkZaZ/dUavVePrpp1FVVRV0dhzNRx99hObm5jF/W7p0KebOneunFk2enp4evPfee2PGxczMzDHHHDt2DBcvXvRH83yC3+3oyhMiH/0akcvl9NFHHzm4Yy+88AL9J7+BpJ6sLzS628LDw+ntt9+mkydPks1mIyIijuOopKSEHnzwQero6KADBw4EpMaCggIyGo00PDwsyMU+f/48aTQakslkk7KtVPp0Oh11dnbS7t27SafT0bFjxwTp3LFjh12jNzqltKFMJiOlUklZWVlUVFREZ86c8ciGwaCxsLCQ7ty5Y+9fE9HY2EirVq2inJycoNHI25KfAVi2bBlt3ryZjEbjhHaczHcK0CgaHMcRx3Fks9mouLjYoS1vvfUWcRwn5lcS+WnMCLRxUcwLGkh2lGSGh2flypVYt24d5s+f77Bv/fr1WLRoEV5++WUHDzBU+PGPf4zi4mLMmTMHsbGxkMv/GzPe1dWFixcv4plnnoHBYPBjK10jk8kQFhY2pt3umD59Ovbv3w+bzQYiwpYtW3Du3Dkft9J75HI5IiIisGzZMkyfPh25ubmCPldcXGz/ZdbT04NNmzbh9u3bvmyq12RlZWHr1q2IjY1FVFQUZs6c6fb40TYEACJCT08PNm/eHLAagRFbCr1Pk5OT8ctf/hL9/f3o7+8HEPh2nD59Ol577TWo1WoAI7OTMTExiIyMdHn8/v370d7ejhs3buAf//gHmpqapGyyRzQ1NeGVV16B1WoFAKfPjXfffRcnT57E1q1bHWZGgolQHhcDzo6uPCHygSdbWlrq1iUbGBggnU4nqScrtkZXm1wup7KyMqe6OY5z6vkGikaZTEZarZaKiopocHDQc1/7PxqfffZZSk5O9viddDDo42lvb6ecnByPYtEm0hdIGjmOo5aWFkpJSQnJ+5Snvb2dZs6cSZGRkQGlkd8KCwvJZDJ5rOvmzZv0+eef04oVK3zSFyd10Wnk/urq6qJPPvmEwsPDJ2yPRqOhCxcuOJynp6eHOjo6BM/yjULyMSMQx0VPL9p4AtWObFm6BCiVSsTGxtp/jQUb4eHh+Pvf/469e/c6xO4IRSaTYdu2bfjkk08wdepUkVs4OcTQxxMXF4ff//73WLdunUitEwexNMpkQrLS+4eIiAjR7BgdHY0NGzZgzZo1IrUuMEhNTcXixYuxZ8+egOyLZrMZ3//+9/H000/DbDZ7dQ4iwqZNm1BUVAS9Xi9yCxlCCFQ7SvpK63+V+Ph45OfnIykpyd9N8Zi8vDzMnj0b06dPn/TDMT4+HkSEFStW4MqVKzh79qxIrfQeMfUBI4GwM2bMgNlsxqpVq1BTU+P3qWixNWo0GixfvhwtLS1obW1FZ2en31/95OXlIScnRzSNSqUSubm5MBqNmDJlCgYGBrx+cItJWFgYFi1ahMLCQofUEEI/HxYWBrVaDZlMhu9+97u4dOkSzp8/74PWekZNTQ3q6urQ2NgoaIDj7+u4uDiHfQaDATdv3sThw4eRk5ODBQsW+KLJDCcEtB1dTf3Q/8jUndganW3f+c536LPPPqObN2861R3Ir7TefvttT6cSBXHw4EHRNAaiPp4NGzaErA3r6+vpd7/7HS1cuDBkNR4/fpzuv/9+Qa/wfK0RGJn6r6ysFC3I02q10v79+8V83nhNSUmJKDYf/zxdvXq1J81gr7REsKMngdZS2jFgXmkdOHAAmzdvRltbm7+bIjrR0dGYNWsW4uPjHfa1tLTg6NGj6O7u9kPLXJOXl4e3334bDzzwgE/O7+9XI57qq6mpwQsvvIATJ0549D3+1OlrjZ2dnThy5AiuX78+iVZKi6cah4aGoNfrA2J257HHHsObb76J9PR0t/dVbW0tSktLHTTyS9VH48kiBF/B2+T48eOCjnd3Xx87dgwvvPACamtr7X8brzmYCNZxUcg194cdJX2lNTQ0BJPJBLVa7TAde+bMGezZswcWi0XKJvkUmUyGiIgIxMfHY9q0aU6P6ejowJkzZ/z+SmA8mZmZKCkpmfA4IoLZbHZ7MyqVSodcNWFhYYiMjITFYvFLbhBP9dXV1aGsrAx333035s2bB2DEvkql0u2KoLCwMCiVSgwPD0v+4BVDI48zG/b19aG2ttavzgD/emaiXEjuNMrlcrsdxzsSHMdheHjYr9nfeY0PPvggnnvuOafHWCwW+0q6uro67Ny5E9OmTbNr5DgOHMdBpVKNiW/itfuzLzY3N6OsrMzpPoVCAZVKNeZvs2fPdnlfV1dXY+fOnWP+ZrPZ0N/fj/DwcHsW6kAiVMZFm80Gi8XiNJdewNjR1dQP+WDqTqvVUkFBAR0/ftxh/qmlpYVOnz5NM2fOlHTqTmyN4/V+/vnn1Nzc7HLebdeuXZSamiookl1KjatWrRI0b9jV1UXLli0jnU7ncvvjH//o8Dmj0UgXLlyg733ve5PW6Et9nZ2dtGzZMsrIyCAAlJaWZtdVVFRE+/bto+rqapef379/Pz3//POk1WoD3obONE5kw6qqKlFs6K3GpUuXUmVlJXV1dXltx+LiYvrss8+osbHR4XN9fX1UW1tLP/jBD0ihUPhF45IlS9xq5DiONm7caNczXuO8efNo3rx5Lu1oMBjo7Nmz9Pjjj4thR48pLy93+X1PPvkkXbhwYczm7nm6fft2h3PExsaSTqejffv2CWmO5GNGII6LQm03mi+//JJ0Oh1NnTo1YO0oqbvb3d0NvV6P3t5eh31paWkOmVCDkcTERKSlpUEmkyExMRH5+flOgyhNJhPq6+tRV1eH9vZ2P7TUOXw9nqysLJfHdHV1oaWlBdOnT4fFYsGlS5fQ0dHh8vjZs2ejsrIS06dPR2xsLAAgNjYW+fn50Gq1omtwhyf6iAhdXV2oqqqyB9+1trba62bdunULc+bMQVRUlMtMoVqtFt/61rccft34Ek808v93pZFnyZIlDueIjY2FTqeT3IbAfzUWFBSgoKDA6TFmsxkDAwPo7OzE9evXXWrU6/WorKxEf38/bDYb7rrrLkRFRQEYeR2dm5uLKVOmSCPMCXFxccjPz3f6Gqurqws3b95EZWUlqqurx+zjNfKzBjabzWlf5M+fmJjoezGjGBoaQl1dHa5du+bymKSkJOTn5ws+Z1JSEgoLC9HU1GQfZ3p7e1FdXR1wYQM8wT4u8nZ0dg/yBIwd3Xl5ENmTBUZyZTjLKEkUGkHLzz//PBmNRrpz547bXCAXLlyg+Ph4UiqVAaWRr6tksVhctr2srIw0Gg0dPnxYUE4WhUJhP348QgIV/aEvIiKCIiIi3M688VmLN2/e7PJctbW1dODAAcrMzAw4GwrRyG/uAivFsKEvNNbV1dGePXtowYIFE9pRpVJRRkYGLVu2jCoqKgJCI7+5m6nj7RgWFiboXL7uiy4b6oS2tjbKyMgglUrl1X3njKGhIRoYGKCioiKHcwkMbJd8zODvwUAaFz255sFkx4B5oXns2DGcO3cuYL1woSiVSkRFRTmNBxgNx3Ewm80BVztMJpPZK9e7Ynh4GCaTCeXl5fj6668xMDDg9pzDw8Ow2Wz2GIPRLFmyBBzH4Z///KfPbZ+VlYUZM2YgOjrarb6srCysXLkSJ0+edDv7RkQYGhpyqotn6tSpsFqtks7wCLWhWBW0lyxZAiLChx9+KEn/5TN+R0REuNUYHx+PmTNnQqlUuo0zIiJYrVbcvn0b33zzjdOK6VLepzyRkZF46qmnsHDhQpfH5ObmYt26dfj4448FBY+764tSw3EcLBaLPQuvGCgUCoSFhTmN8fjyyy8hk8nwxBNP+GVW0huCYVwMJjsGjMNz8OBBl4FrwYRcLvcqP0awQP/9lYPdu3dP+nxr1qzB8uXLcerUKZ936ry8PBQUFLhNAElE0Ol0iI2NRVtb26RfNyYnJyMmJmbSifDEZLQNxYC34VdffSXJg1lI2QgiQmJiIrRardPVkc7o7e3FnTt3nDrwUt6nPHFxcfjtb3+LlJQUp/s5jsP999+Pb3/722hubg6q1XJCkMlkoq5yLC8vx2effYb77rsvaByeUBgXA8mOAePwMAIfvi5KfX29v5viFZmZmcjNzXW5oofXd+fOHZhMJly5ckXiFvoeb23Y29uL5uZmJCcnQ6PR+Kh1wuA4zl6fzRlC6ve4QkxH0BcQEbq7u9HU1IQ33ngDQ0NDICJREgeq1WpERkZicHDQr6vSgJHaX3/4wx8mrPXGCGwCzY7M4ZEYIoJer0dXV1fAPVwTEhKQnJzsMEPFt7mxsRGHDx8OiJwkniKTyTB16lSkpqa6nB3o7e3FJ598ItqrntHfPWXKFMTHx/s8/YArG/J4q9FoNKKpqQnR0dF+d3ji4+ORlJTkVCOfN8fb+9TfOWl43PXF1tZWXL58GYcPHxb1NYJcLodCoZAkd5TBYEBnZ6fDqzW+n86YMQMrVqwIqJlRhiPBZkfm8EgMX2OkqqoqoBwHvtZVUVGRw6qyQG2zp2RnZ0On04keT8O/InI1UKjVamzduhWVlZUoKSkRdZAajTsbTpZPP/0Up06dwt69e/Hwww+Lem5PcKeR4zjU19ejvr7eqx8TKpUKERERbnN88LEFvoyBcafRYrHgZz/7GaqqqkS/j8xmM/r6+nz+Q4xopEbSkSNHHEoP8DXf8vPzA2aQZDgnGO3IHB4Rmehdpac1RqQmISHB6dJUfhrd2zZnZWUhLy8PycnJDvv4a2I0Gr06txAyMjIwZ84cpKWlOXV2bDYbTp48ifr6ehQWFqKlpQU3btwQ7fs5jkNjYyMaGxt9Ppi4siGvsbKy0qvBOjU1FXl5eUhISBCjmV7BJ+DTarVONXIch3PnzuHSpUtIT0+HwWCAwWAQfP7MzEzMmzdvwvtUiplZd33x9u3bXvcXd31R7NgudxgMBnR1dTn8nZ8Z8NZhv3btGmpqapymyXBXs4nhHcFmR+bwiMhE0+G7d+/Gzp07A+5Vlq955JFHsGPHDqf7du/e7fOgvIcffhhbtmxBTEyM0/0WiwWbN29GX18fXnvtNXz66ad47733BJ9/IkfXarVi69atLnNUSAGv8cKFC1593p0NAUgS88FnLXc1Q2e1WrFz5060tbVh1apVuHLlCk6dOiX4/A8//LBf71MpeOSRR7Bt2zanz6pQeC4dOnQIpaWlTvetW7dOUOZxhv/xlR2ZwyMS06ZNw4svvuiQkn88ofBQ8RR/183iUwW4Wz3HcRy6u7vxl7/8RfDsjkKhQFxcnD1JnTOsVisGBgZ8bvewsDCoVKoJNXqLOxseOHAAR48edUhWKDYWiwW9vb0TpnLo7+/H6dOnPZrdAdxrVKvViIqKgslk8ltAb19fH/R6/aTKPxw/fhwvvvgi1q9fj9zc3DH7/N1PxeB/8fkaivjKjszhEQmtVouf/OQnAZ0R0xv4Oi/e3IByuRxqtVrSHDTOUCgUbpei8/T29uLIkSOCz6tUKu3Lzl1hMplw+/Ztn9coUqvViI2NFb1WkBAbnjhxAu+8846o3+sMPufRRNdycHAQly9fFnxeIRqVSiXCw8NhNpv95vAYjUa0tbV5nbsrLCzMntl4+fLldofHXQ0kKVEqldBoNF4FjvsiFwzDOwLZjszhYbhl+/bt+OCDD3D16lWPP5udnY2//vWvSEtL80HL/E9mZib27t2L1NRUl8e88847eP/9992mzxeD1atX46WXXkJGRoao5w11GwLCNJpMJhiNRr8U1+TZu3cv3n//fa/y7cTExGD+/PlYvHgxiouLkZmZCWDkl/SJEyfw0ksv4ebNmyK32DNefPFFPPXUU8jOzvb4sw0NDfjRj35kL5fC8B+BbEdJHZ7ExERMmzbNabBReno6dDodGhoaRF8W7Evkcjmys7ORk5Pj0qM1Go24du0a1Gp10GlsbW3FxYsXJzwuLi4OWVlZaGlpgV6vR3Z2Nu69917k5+c7nV3hr4mzgDd/EB4eDrVaLagqMW/zgoIC5OTkuNVXXV3t0WyDt/BVvcWaCuY1BpMN5XI55syZA7PZjIaGhgmvhScab9265VdnBxipm9XQ0ODRZ3iNycnJKCwsRH5+PnQ63Zhjent7UVVV5ffcO2lpaS5r0rmC4zg0NDTg/PnzqKqqctp/+WeT1LXChBJq42JA29FVzQnyQc2QDRs2kMlkIpvN5lD4wmq1kl6vp7lz50paM2Sy3xEREUEVFRVkNptdFvX46KOPSKPR0FtvvRWwGl3VchFSXwcAFRcX08DAAD3//POk0WgEXxMhNYAmq2/jxo3EcZzLtgwODtKjjz5KGRkZJJPJRLW5GPqEaFQqlRQVFeW0RhKR5/V4AlEjAJLL5S5rDhERmc1mOnnypKD6YFLfp0I1TrYvjrfjsWPHyGg0ksVioeHh4THn5DiODh48KOi+F6hxQjiOo+LiYodzC6yRNAaTyUQLFy4ktVo94bNpvHYX+HzMGL8F4rgo5EIFox0lneFRKBQuY1xOnTqFc+fOBeRybXfIZDKo1Wq3MSI2mw0mkwnHjh2D2WwOKo1CawiFhYVBo9HgoYceQnR0NNLS0lxeE4vFArPZ7LRmkS8wGAy4fPky7r77bkRHRzvsVygUWLFiBVJSUrBr1y579trRZGVl4ZFHHoFMJoNKpXKpz2Qy4eDBgzh9+rRk+oD/xrdMJj+MUI1WqxU1NTW4fPmypBrlcjmUSqXL4FoiQltbGzo6OlzO7mRkZKCoqMj+LHKlsb+/H/v27UNFRYWkGt3BV6zny1vcunULNpvN5czMAw88gPnz5yMjI8Np0P5oja6uV6BDRDCbzW5nZvlnU6ASiuOip7iz4+iacpO1o6QOD+9lOXtglZeXo6ysLGg6Hq/B3ZJkXgv/b3l5OcrLy6VpoBc4sw9fQ6iiosJtp+M/s2bNGqxZs8blcRzHwWQySZrA8NatW/j6668RGxvr0uF55plncM899+Bvf/ubvXTBaPLy8twuy+bp6+vDG2+8IWoeH6FM1HcmKmgrVKPZbMYXX3whSjkDT5hoJRrHcbhy5Yq9JIgzrbm5ufi///s/aDQat0HKRqMRv/71r53mAZECV31xxYoV+NWvfoXq6mr09va6DfB89NFHsWHDBqfXgYj8qpF/bnr6vB9//ESv4Sa65wOBYB4XpbDjRDXlPEFSh+fjjz/GN998g1dffRXz588fs2/9+vVYtGgRXn75ZTQ3N0vZLI9ZuHAhfvGLX9irNvMBgKPZunUrzp8/D6vVOukClFJARNiyZQvKy8uxbdu2MQUX1Wo1tm/fjt7eXpefd5bIzBm3b99GSUkJ6urqJt1moVRXV6O7uxtZWVlIT093eoxcLseMGTOwb98+p3WahOrjOA5Go3HCCvJSI6YNBwcH7X1ZSoaHh2E2m+2xSuMHCLlcjnvvvRfZ2dkoLCx0+hBOSkqaMEWB1WqFxWLx2yDjri8qlUr88Ic/RG9vL4xGo9uaYrm5uS4H+y1btuDEiRMeL90XA5lMhldffRWrV6/Gpk2bBJdbMRgM2LRp05g2Dw8PuxwvEhISsH37duTk5IjSbl8RrOOiVHYUE0kdns7OTlitVqcP3dzcXGRmZmLLli1SNskrkpOTsXLlSpf7iUaK+R06dCioSjGcO3cObW1tDm0OCwvD4sWLvT5vf3+/PTNsZ2cnjh49KmmgK58l2t1gDwCxsbEoKiry+PxEZHd0Ojo6MDg46LclvnxtG61WOyaIXmwbNjU1SR6sTEQYHh6GXq9HR0cHkpOTx2iUyWRISkpCUlKSVytEeI0WiwVtbW0+LR8xEa76olwun9QAzms8ceIE/v3vf0+2mV4zf/58pKamCi47YDAY8M033+DIkSNu77vIyEh78G9ycjKWL18esMHKPME8LkplR3c/UDzCVXAP+SA4a/369dTe3u4ySNDTwMqJNnfaJqNx1apVbqOlOI6jlStXCg4EDCSNd911F7W3t7vV5ym7du2ilJQUSklJocTERJLL5aJqFHIOV4GgYnDnzh1qaWmhJ554wif6hGoEQElJSbRw4ULq6uoSVaOvbeiJxvj4eNLpdNTZ2ekTjcnJyaTVav2q0dd90V1w6CQ1CqalpYVSUlLs53YV7MpxHD3zzDOC7jt+jGlvb6fOzk6ngcATIHnQciCOi55csGCyo6QzPJGRkaK8h/M3HR0dKC8vx9y5c52+zgIwunMEFYODg/j000+Rk5ODBQsWeHUOvm6TTCbDokWLMDAw4LdYCB4isgdnJiQkYOrUqZg1a5Yo575+/ToqKipQX1/v9yXaPT09kMlkOHTo0KRsOJ5AsCFPb28v2tvbcfjwYcyZM8fhNYAn3LlzB0ajEVVVVaioqAgYjYC4fZGfnQskjRqNBsuXL8eVK1dw9uxZp8fwNZUuX77stG+FhYVh0aJFiIyMRE9PD3JycoJujAn2cVFMO/KzOgsXLvTNNXHn5UFkT7a0tNStSxYsMzz8VlZW5tKTdbZczxebrzSuXr3aU496jB3nzZtHCxYsIJPJRNu3b/epRk/OFR4eTg888AD95je/8VrfeHbs2BFyNhyPr23oicbw8HDSaDQkl8snpdFms1FtbS3t2bNH0DL2YLTjwMAAFRQUTFqbBxo95uDBgwQ4nxnYsWOH25lyjUZDlZWVZDAY6F//+hfV1NR4e6l4JJ/hCcRx0ZsLN5Ed3bWJt6OI+H+G58svv8TGjRud1nEJRmjUDM7g4CCuXr2K2tpanD17FrW1tX5s2eS5ePEiSkpKxgQ9pqeno7S0FEql0u1nVSoVfv7zn6OlpQWvvPKK1wUrfcHQ0BAaGxsxODgIvV6Pxx9/3KPYlgMHDuDEiRNj/lZVVSV2M0VhtA1VKhVKS0sFZUwOdI1DQ0P21SG1tbUoKSnB0qVLkZ+fj+Tk5AlrigEjGo8fP46enh50d3f7vayCO5z1xWnTpmHjxo1O+6LNZsPg4CAOHTqE48ePB3z24blz56KsrMxpP1yyZAneeust7Nq1C5cuXRqz77HHHsODDz6I9PR0REZG4p577nFb5iVQcTcufvHFFzhz5ozfZ46FIMSO7777rls7+hxXnhD5wJMFRmIp9u/fTwMDAzQwMGB/b2mxWKirqysgk/K52t544w27jra2Nvrggw9o7dq1ol8zf2ocvel0OtLr9XbNozfejlar1f43oQngJqtRLBsK2Z577rmgtGFkZCSdPn06pDQqFAp7QsCf/vSndPToUWprayOj0UgDAwNkMpnIbDaTyWQKWo2uNnd98fbt29TS0kLPPvusPzSKDsdx9Nhjj5FGoxmz/elPf/LF10k+wwO4Hhdfe+01ys3NFeU56sl96osLGwh2lNGoWYrx/GcqUXQyMzPtnvjSpUuxbds2vP766/aaTWKl0CaiCRMwTEZjWloapk6dCmBkWV1/fz/6+vokXerpa42jiYiIQHZ2ttOlrqPtuG/fPgD/nfVyd48JYSKNYtlQCG1tbW4TMHqDFDbkl90LKW4bjBq1Wi3i4+OhUqnsBVRTUlJw9913o6GhwWHJbDBqHI1Go8GMGTNcLjsfHh5GR0cHenp6xPg6OwI0+mTMaG5udljFlJaWBq1WK/ZXSWbD8YweFxcvXozf//73ePXVV/Huu++KmiZByBCnlf4AAAFNSURBVH2KELWjX4qHjl5vHxcXh/Pnz+PChQuCajYFEq2trWhtbfV3MyRjcHDQpY1G27G6ulrilnnP/4oN+Vo1oUp3d7eDA3Pr1i309PTg6tWr9qDdUMFkMgXd83IyuFocEkqMHhejoqJQWVmJ1tbWoEptMhH+tqNfZnhGI5fLoVarYbVaRc97IeUvLn8RKBr9aUdmw8DHHxr5OB+pimIyO44cIklDfEdA2FAul0OlUtlLxoiJP2d4JMSpRr87PL6EPYBGCHWNoa4PYBqDAaZx5BBJGuI7mA3/c5jPG+JbnGqUO/sjg8FgMBgMRijBHB4Gg8FgMBghD3N4GAwGg8FghDxuY3gYDAaDwWAwQgE2w8NgMBgMBiPkYQ4Pg8FgMBiMkIc5PAwGg8FgMEIe5vAwGAwGg8EIeZjDw2AwGAwGI+RhDg+DwWAwGIyQ5/8BsWlK+WKaMyAAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 720x7200 with 10 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch: 1000 of 1000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "e0EwkI11i5-o"
      },
      "source": [
        "In order to generate the figures the following code can be employed."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ksJsjVMV17UL",
        "outputId": "a06d4e12-a766-4389-d2cf-1f8117852419",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 79
        }
      },
      "source": [
        "n_examples = 5\n",
        "scale_image = 5\n",
        "noise= np.random.normal(loc = mu, scale = sigma, size = (n_examples, input_dim))\n",
        "generated_images = generator_model.predict(noise)\n",
        "\n",
        "fig, axes = plt.subplots(1,n_examples,\n",
        "                         figsize = (scale_image, scale_image * n_examples))\n",
        "for i in range(generated_images.shape[0]):\n",
        "    axes[i].imshow(generated_images[i], cmap = \"gray\")\n",
        "    axes[i].axis('off')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x1800 with 5 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JOJlh6BewLiS"
      },
      "source": [
        "## License\n",
        "\n",
        "This Deep Learning Tutorial is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 (CC BY-NC-ND 4.0) International License."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OoGgQqWDvSs9"
      },
      "source": [
        "## Acknowledgments\n",
        "Henrique F. de Arruda acknowledges FAPESP for sponsorship (grant no. 2018/10489-0). H. F. de Arruda also thanks Soremartec S.A. and Soremartec Italia, Ferrero Group, for partial financial support (from 1st July 2021). His funders had no role in study design, data collection, and analysis, decision to publish, or manuscript preparation. Alexandre Benatti thanks Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - Brasil (CAPES) - Finance Code 001. Luciano da F. Costa thanks CNPq (grant no. 307085/2018-0) and FAPESP (proc. 15/22308-2) for sponsorship. César H. Comin thanks FAPESP (Grant Nos. 2018/09125-4 and 2021/12354-8) for financial support. This work has been supported also by FAPESP grants 11/50761-2 and 15/22308-2."
      ]
    }
  ]
}