ML4DM_Lab_2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "ML4DM_Lab_2.ipynb",
      "version": "0.3.2",
      "provenance": [],
      "collapsed_sections": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/nishi1612/Machine-Learning-For-Data-Mining/blob/master/ML4DM_Lab_2.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kxPwxyPfJN_7",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        },
        "outputId": "e3bd9734-c7c7-4d04-967e-582b9792db0a"
      },
      "source": [
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "from keras.datasets import fashion_mnist\n",
        "import keras\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "fwI89YbGJjxF",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "((trainX, trainY), (testX, testY)) = fashion_mnist.load_data()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qJCkkIhZJsNV",
        "colab_type": "code",
        "outputId": "6f9db37d-e412-4de5-d4c0-db71e7cdc7f4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "trainX.shape , trainY.shape, testX.shape, testY.shape"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "((60000, 28, 28), (60000,), (10000, 28, 28), (10000,))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LoV5CeF0La9X",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "trainX = trainX/255.0\n",
        "testX = testX/255.0"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "bG51GmB-sPrh",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "def show_graph(history):\n",
        "  plt.plot(history.history['loss'])\n",
        "  plt.title('Loss')\n",
        "  plt.xlabel('No. of epochs')\n",
        "  plt.ylabel('Loss')\n",
        "  plt.show()\n",
        "  plt.plot(history.history['acc'])\n",
        "  plt.title('Accuracy')\n",
        "  plt.xlabel('No. of epochs')\n",
        "  plt.ylabel('Accuracy')\n",
        "  plt.show()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HqM7LrLNtWwX",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "epochs = 100\n",
        "verbose = 1\n",
        "validation_split = 0.2"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BOVT3D0NLjDA",
        "colab_type": "code",
        "outputId": "bc62a04a-6366-40c8-f7fd-53c2b03374c4",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 445
        }
      },
      "source": [
        "model = keras.Sequential([\n",
        "    keras.layers.Flatten(input_shape=(trainX.shape[1],trainX.shape[2])),\n",
        "    keras.layers.Dense(256,activation = tf.nn.relu),\n",
        "    keras.layers.Dense(10,activation=tf.nn.softmax)    \n",
        "])\n",
        "model.compile(optimizer='adam',\n",
        "              loss='sparse_categorical_crossentropy',\n",
        "              metrics=['accuracy'])\n",
        "model.summary()"
      ],
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:66: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:541: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4432: The name tf.random_uniform is deprecated. Please use tf.random.uniform instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/optimizers.py:793: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n",
            "\n",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3622: The name tf.log is deprecated. Please use tf.math.log instead.\n",
            "\n",
            "Model: \"sequential_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "flatten_1 (Flatten)          (None, 784)               0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 256)               200960    \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 10)                2570      \n",
            "=================================================================\n",
            "Total params: 203,530\n",
            "Trainable params: 203,530\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1syPRIZjNXc5",
        "colab_type": "code",
        "outputId": "efebfc8a-c342-405f-9258-94570670a632",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "history = model.fit(trainX, trainY, epochs=epochs, validation_split=validation_split, verbose=verbose, shuffle = True)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/math_grad.py:1250: add_dispatch_support.<locals>.wrapper (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",
            "WARNING:tensorflow:From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:1033: The name tf.assign_add is deprecated. Please use tf.compat.v1.assign_add instead.\n",
            "\n",
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/100\n",
            "48000/48000 [==============================] - 7s 148us/step - loss: 0.5047 - acc: 0.8216 - val_loss: 0.4331 - val_acc: 0.8506\n",
            "Epoch 2/100\n",
            " 1120/48000 [..............................] - ETA: 12s - loss: 0.3930 - acc: 0.8607"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "/usr/local/lib/python3.6/dist-packages/keras/callbacks.py:95: RuntimeWarning: Method (on_train_batch_end) is slow compared to the batch update (0.125727). Check your callbacks.\n",
            "  % (hook_name, delta_t_median), RuntimeWarning)\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "stream",
          "text": [
            "48000/48000 [==============================] - 7s 145us/step - loss: 0.3769 - acc: 0.8622 - val_loss: 0.3650 - val_acc: 0.8672\n",
            "Epoch 3/100\n",
            "48000/48000 [==============================] - 7s 140us/step - loss: 0.3363 - acc: 0.8779 - val_loss: 0.3650 - val_acc: 0.8662\n",
            "Epoch 4/100\n",
            "48000/48000 [==============================] - 7s 142us/step - loss: 0.3096 - acc: 0.8858 - val_loss: 0.3321 - val_acc: 0.8807\n",
            "Epoch 5/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.2902 - acc: 0.8923 - val_loss: 0.3243 - val_acc: 0.8834\n",
            "Epoch 6/100\n",
            "48000/48000 [==============================] - 7s 154us/step - loss: 0.2740 - acc: 0.8994 - val_loss: 0.3299 - val_acc: 0.8848\n",
            "Epoch 7/100\n",
            "48000/48000 [==============================] - 7s 144us/step - loss: 0.2614 - acc: 0.9021 - val_loss: 0.3202 - val_acc: 0.8847\n",
            "Epoch 8/100\n",
            "48000/48000 [==============================] - 7s 141us/step - loss: 0.2512 - acc: 0.9070 - val_loss: 0.3105 - val_acc: 0.8892\n",
            "Epoch 9/100\n",
            "48000/48000 [==============================] - 7s 141us/step - loss: 0.2386 - acc: 0.9124 - val_loss: 0.3181 - val_acc: 0.8874\n",
            "Epoch 10/100\n",
            "48000/48000 [==============================] - 7s 143us/step - loss: 0.2303 - acc: 0.9144 - val_loss: 0.3150 - val_acc: 0.8901\n",
            "Epoch 11/100\n",
            "48000/48000 [==============================] - 7s 145us/step - loss: 0.2217 - acc: 0.9170 - val_loss: 0.3341 - val_acc: 0.8859\n",
            "Epoch 12/100\n",
            "48000/48000 [==============================] - 7s 146us/step - loss: 0.2135 - acc: 0.9201 - val_loss: 0.3212 - val_acc: 0.8860\n",
            "Epoch 13/100\n",
            "48000/48000 [==============================] - 7s 145us/step - loss: 0.2036 - acc: 0.9236 - val_loss: 0.3572 - val_acc: 0.8828\n",
            "Epoch 14/100\n",
            "48000/48000 [==============================] - 7s 143us/step - loss: 0.1982 - acc: 0.9263 - val_loss: 0.3182 - val_acc: 0.8906\n",
            "Epoch 15/100\n",
            "48000/48000 [==============================] - 7s 142us/step - loss: 0.1918 - acc: 0.9286 - val_loss: 0.3257 - val_acc: 0.8901\n",
            "Epoch 16/100\n",
            "48000/48000 [==============================] - 7s 142us/step - loss: 0.1866 - acc: 0.9294 - val_loss: 0.3367 - val_acc: 0.8890\n",
            "Epoch 17/100\n",
            "48000/48000 [==============================] - 7s 143us/step - loss: 0.1786 - acc: 0.9333 - val_loss: 0.3401 - val_acc: 0.8905\n",
            "Epoch 18/100\n",
            "48000/48000 [==============================] - 7s 140us/step - loss: 0.1729 - acc: 0.9345 - val_loss: 0.3429 - val_acc: 0.8922\n",
            "Epoch 19/100\n",
            "48000/48000 [==============================] - 7s 142us/step - loss: 0.1650 - acc: 0.9382 - val_loss: 0.3334 - val_acc: 0.8972\n",
            "Epoch 20/100\n",
            "48000/48000 [==============================] - 7s 139us/step - loss: 0.1631 - acc: 0.9380 - val_loss: 0.3379 - val_acc: 0.8920\n",
            "Epoch 21/100\n",
            "48000/48000 [==============================] - 7s 142us/step - loss: 0.1580 - acc: 0.9406 - val_loss: 0.3345 - val_acc: 0.8905\n",
            "Epoch 22/100\n",
            "48000/48000 [==============================] - 7s 142us/step - loss: 0.1518 - acc: 0.9436 - val_loss: 0.3552 - val_acc: 0.8882\n",
            "Epoch 23/100\n",
            "48000/48000 [==============================] - 7s 142us/step - loss: 0.1487 - acc: 0.9441 - val_loss: 0.3510 - val_acc: 0.8939\n",
            "Epoch 24/100\n",
            "48000/48000 [==============================] - 7s 141us/step - loss: 0.1414 - acc: 0.9475 - val_loss: 0.3618 - val_acc: 0.8921\n",
            "Epoch 25/100\n",
            "48000/48000 [==============================] - 7s 137us/step - loss: 0.1416 - acc: 0.9472 - val_loss: 0.3549 - val_acc: 0.8929\n",
            "Epoch 26/100\n",
            "48000/48000 [==============================] - 7s 140us/step - loss: 0.1344 - acc: 0.9499 - val_loss: 0.3800 - val_acc: 0.8944\n",
            "Epoch 27/100\n",
            "48000/48000 [==============================] - 7s 140us/step - loss: 0.1311 - acc: 0.9504 - val_loss: 0.3815 - val_acc: 0.8902\n",
            "Epoch 28/100\n",
            "48000/48000 [==============================] - 7s 141us/step - loss: 0.1288 - acc: 0.9523 - val_loss: 0.4152 - val_acc: 0.8826\n",
            "Epoch 29/100\n",
            "48000/48000 [==============================] - 7s 143us/step - loss: 0.1257 - acc: 0.9534 - val_loss: 0.3776 - val_acc: 0.8944\n",
            "Epoch 30/100\n",
            "48000/48000 [==============================] - 7s 146us/step - loss: 0.1218 - acc: 0.9535 - val_loss: 0.3821 - val_acc: 0.8948\n",
            "Epoch 31/100\n",
            "48000/48000 [==============================] - 7s 143us/step - loss: 0.1182 - acc: 0.9549 - val_loss: 0.3898 - val_acc: 0.8965\n",
            "Epoch 32/100\n",
            "48000/48000 [==============================] - 7s 140us/step - loss: 0.1164 - acc: 0.9568 - val_loss: 0.3851 - val_acc: 0.8918\n",
            "Epoch 33/100\n",
            "48000/48000 [==============================] - 7s 138us/step - loss: 0.1135 - acc: 0.9570 - val_loss: 0.4284 - val_acc: 0.8844\n",
            "Epoch 34/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.1109 - acc: 0.9577 - val_loss: 0.4186 - val_acc: 0.8932\n",
            "Epoch 35/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.1067 - acc: 0.9599 - val_loss: 0.4182 - val_acc: 0.8926\n",
            "Epoch 36/100\n",
            "48000/48000 [==============================] - 7s 138us/step - loss: 0.1072 - acc: 0.9596 - val_loss: 0.4362 - val_acc: 0.8914\n",
            "Epoch 37/100\n",
            "48000/48000 [==============================] - 7s 146us/step - loss: 0.1041 - acc: 0.9607 - val_loss: 0.4467 - val_acc: 0.8889\n",
            "Epoch 38/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.1007 - acc: 0.9628 - val_loss: 0.4457 - val_acc: 0.8883\n",
            "Epoch 39/100\n",
            "48000/48000 [==============================] - 8s 167us/step - loss: 0.1012 - acc: 0.9613 - val_loss: 0.4393 - val_acc: 0.8929\n",
            "Epoch 40/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.0955 - acc: 0.9632 - val_loss: 0.4414 - val_acc: 0.8888\n",
            "Epoch 41/100\n",
            "48000/48000 [==============================] - 7s 137us/step - loss: 0.0952 - acc: 0.9636 - val_loss: 0.4686 - val_acc: 0.8872\n",
            "Epoch 42/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0935 - acc: 0.9650 - val_loss: 0.4543 - val_acc: 0.8919\n",
            "Epoch 43/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.0887 - acc: 0.9666 - val_loss: 0.4743 - val_acc: 0.8884\n",
            "Epoch 44/100\n",
            "48000/48000 [==============================] - 7s 146us/step - loss: 0.0918 - acc: 0.9662 - val_loss: 0.4738 - val_acc: 0.8871\n",
            "Epoch 45/100\n",
            "48000/48000 [==============================] - 7s 145us/step - loss: 0.0877 - acc: 0.9664 - val_loss: 0.4517 - val_acc: 0.8946\n",
            "Epoch 46/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.0854 - acc: 0.9685 - val_loss: 0.4526 - val_acc: 0.8946\n",
            "Epoch 47/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0839 - acc: 0.9682 - val_loss: 0.5181 - val_acc: 0.8850\n",
            "Epoch 48/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0800 - acc: 0.9700 - val_loss: 0.4687 - val_acc: 0.8916\n",
            "Epoch 49/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0818 - acc: 0.9683 - val_loss: 0.4802 - val_acc: 0.8949\n",
            "Epoch 50/100\n",
            "48000/48000 [==============================] - 7s 137us/step - loss: 0.0775 - acc: 0.9713 - val_loss: 0.4820 - val_acc: 0.8937\n",
            "Epoch 51/100\n",
            "48000/48000 [==============================] - 7s 144us/step - loss: 0.0777 - acc: 0.9711 - val_loss: 0.5070 - val_acc: 0.8925\n",
            "Epoch 52/100\n",
            "48000/48000 [==============================] - 7s 139us/step - loss: 0.0748 - acc: 0.9714 - val_loss: 0.4869 - val_acc: 0.8928\n",
            "Epoch 53/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0765 - acc: 0.9709 - val_loss: 0.5035 - val_acc: 0.8902\n",
            "Epoch 54/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0752 - acc: 0.9723 - val_loss: 0.5174 - val_acc: 0.8872\n",
            "Epoch 55/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0707 - acc: 0.9738 - val_loss: 0.4994 - val_acc: 0.8962\n",
            "Epoch 56/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0726 - acc: 0.9731 - val_loss: 0.5260 - val_acc: 0.8901\n",
            "Epoch 57/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.0687 - acc: 0.9741 - val_loss: 0.5161 - val_acc: 0.8906\n",
            "Epoch 58/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0698 - acc: 0.9739 - val_loss: 0.5583 - val_acc: 0.8908\n",
            "Epoch 59/100\n",
            "48000/48000 [==============================] - 7s 141us/step - loss: 0.0653 - acc: 0.9757 - val_loss: 0.5367 - val_acc: 0.8898\n",
            "Epoch 60/100\n",
            "48000/48000 [==============================] - 7s 146us/step - loss: 0.0650 - acc: 0.9752 - val_loss: 0.5274 - val_acc: 0.8919\n",
            "Epoch 61/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0671 - acc: 0.9752 - val_loss: 0.5258 - val_acc: 0.8904\n",
            "Epoch 62/100\n",
            "48000/48000 [==============================] - 6s 133us/step - loss: 0.0634 - acc: 0.9766 - val_loss: 0.5525 - val_acc: 0.8916\n",
            "Epoch 63/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.0636 - acc: 0.9760 - val_loss: 0.5409 - val_acc: 0.8950\n",
            "Epoch 64/100\n",
            "48000/48000 [==============================] - 7s 139us/step - loss: 0.0607 - acc: 0.9773 - val_loss: 0.5331 - val_acc: 0.8968\n",
            "Epoch 65/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0624 - acc: 0.9766 - val_loss: 0.5611 - val_acc: 0.8877\n",
            "Epoch 66/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0586 - acc: 0.9782 - val_loss: 0.5899 - val_acc: 0.8880\n",
            "Epoch 67/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0600 - acc: 0.9779 - val_loss: 0.5618 - val_acc: 0.8947\n",
            "Epoch 68/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0596 - acc: 0.9768 - val_loss: 0.5653 - val_acc: 0.8891\n",
            "Epoch 69/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0556 - acc: 0.9793 - val_loss: 0.5977 - val_acc: 0.8872\n",
            "Epoch 70/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0592 - acc: 0.9778 - val_loss: 0.5887 - val_acc: 0.8894\n",
            "Epoch 71/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.0546 - acc: 0.9796 - val_loss: 0.5801 - val_acc: 0.8883\n",
            "Epoch 72/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0562 - acc: 0.9787 - val_loss: 0.6030 - val_acc: 0.8918\n",
            "Epoch 73/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0544 - acc: 0.9799 - val_loss: 0.5758 - val_acc: 0.8916\n",
            "Epoch 74/100\n",
            "48000/48000 [==============================] - 7s 136us/step - loss: 0.0529 - acc: 0.9803 - val_loss: 0.6279 - val_acc: 0.8892\n",
            "Epoch 75/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0540 - acc: 0.9800 - val_loss: 0.5903 - val_acc: 0.8932\n",
            "Epoch 76/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0527 - acc: 0.9806 - val_loss: 0.6115 - val_acc: 0.8915\n",
            "Epoch 77/100\n",
            "48000/48000 [==============================] - 7s 135us/step - loss: 0.0543 - acc: 0.9799 - val_loss: 0.6227 - val_acc: 0.8910\n",
            "Epoch 78/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0488 - acc: 0.9819 - val_loss: 0.6412 - val_acc: 0.8880\n",
            "Epoch 79/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0520 - acc: 0.9800 - val_loss: 0.6059 - val_acc: 0.8948\n",
            "Epoch 80/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0488 - acc: 0.9819 - val_loss: 0.6141 - val_acc: 0.8913\n",
            "Epoch 81/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0481 - acc: 0.9823 - val_loss: 0.6392 - val_acc: 0.8889\n",
            "Epoch 82/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0480 - acc: 0.9823 - val_loss: 0.6313 - val_acc: 0.8878\n",
            "Epoch 83/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.0492 - acc: 0.9816 - val_loss: 0.6447 - val_acc: 0.8919\n",
            "Epoch 84/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.0471 - acc: 0.9823 - val_loss: 0.6297 - val_acc: 0.8924\n",
            "Epoch 85/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0473 - acc: 0.9824 - val_loss: 0.6406 - val_acc: 0.8943\n",
            "Epoch 86/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0472 - acc: 0.9828 - val_loss: 0.6392 - val_acc: 0.8915\n",
            "Epoch 87/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0460 - acc: 0.9827 - val_loss: 0.6726 - val_acc: 0.8919\n",
            "Epoch 88/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.0432 - acc: 0.9840 - val_loss: 0.6627 - val_acc: 0.8872\n",
            "Epoch 89/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.0431 - acc: 0.9841 - val_loss: 0.6592 - val_acc: 0.8937\n",
            "Epoch 90/100\n",
            "48000/48000 [==============================] - 6s 133us/step - loss: 0.0490 - acc: 0.9821 - val_loss: 0.6689 - val_acc: 0.8912\n",
            "Epoch 91/100\n",
            "48000/48000 [==============================] - 8s 159us/step - loss: 0.0402 - acc: 0.9851 - val_loss: 0.6590 - val_acc: 0.8877\n",
            "Epoch 92/100\n",
            "48000/48000 [==============================] - 7s 142us/step - loss: 0.0435 - acc: 0.9842 - val_loss: 0.6597 - val_acc: 0.8928\n",
            "Epoch 93/100\n",
            "48000/48000 [==============================] - 7s 145us/step - loss: 0.0418 - acc: 0.9846 - val_loss: 0.6496 - val_acc: 0.8932\n",
            "Epoch 94/100\n",
            "48000/48000 [==============================] - 6s 133us/step - loss: 0.0414 - acc: 0.9850 - val_loss: 0.6784 - val_acc: 0.8919\n",
            "Epoch 95/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0421 - acc: 0.9845 - val_loss: 0.6483 - val_acc: 0.8927\n",
            "Epoch 96/100\n",
            "48000/48000 [==============================] - 6s 135us/step - loss: 0.0379 - acc: 0.9857 - val_loss: 0.6646 - val_acc: 0.8899\n",
            "Epoch 97/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.0467 - acc: 0.9831 - val_loss: 0.7299 - val_acc: 0.8879\n",
            "Epoch 98/100\n",
            "48000/48000 [==============================] - 7s 138us/step - loss: 0.0410 - acc: 0.9852 - val_loss: 0.6907 - val_acc: 0.8922\n",
            "Epoch 99/100\n",
            "48000/48000 [==============================] - 7s 144us/step - loss: 0.0382 - acc: 0.9856 - val_loss: 0.6665 - val_acc: 0.8917\n",
            "Epoch 100/100\n",
            "48000/48000 [==============================] - 6s 134us/step - loss: 0.0375 - acc: 0.9863 - val_loss: 0.6660 - val_acc: 0.8912\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Gq20HgHRZWdA",
        "colab_type": "code",
        "outputId": "336ebdde-44d2-4d66-ab08-6ac250b199fd",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        }
      },
      "source": [
        "show_graph(history)"
      ],
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4zLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvnQurowAAIABJREFUeJzt3Xd4HOW5/vHvo1WzqmVbsmXJ3cbG\nuFuYEnogARIMAUIJhEDIjzRCTgon5CQhOeTil0JI5xAIOJQTejUlQBJabMBYLhhwQ+4yLpIsS7Ks\nvs/5Y8eKsCVZLquVtPfnuvbCMzu784zH6Na878z7mrsjIiICkBDrAkREpOdQKIiISCuFgoiItFIo\niIhIK4WCiIi0UiiIiEgrhYKIiLRSKIh0wMzWm9npsa5DpDspFEREpJVCQeQAmdn/M7MSM9thZnPN\nbGiw3szsN2a23cyqzexdM5sUvHe2mS03sxoz22xm343tUYi0T6EgcgDM7DTgZ8BFQD6wAXgoePsT\nwEnAEUB2sE1F8N7dwJfdPROYBLzcjWWLdFlirAsQ6WUuA+a4+2IAM/s+UGlmI4EmIBOYALzt7iva\nfK4JmGhm77h7JVDZrVWLdJGuFEQOzFAiVwcAuPsuIlcDBe7+MvBH4DZgu5ndaWZZwaYXAGcDG8zs\nNTM7rpvrFukShYLIgfkQGLFnwczSgYHAZgB3/727zwQmEmlGuj5Yv9DdzwXygKeAR7q5bpEuUSiI\ndC7JzFL3vIAHgavMbJqZpQD/H1jg7uvN7GgzO8bMkoBaoB4Im1mymV1mZtnu3gRUA+GYHZFIJxQK\nIp17Hqhr8zoF+BHwOLAFGANcEmybBfyZSH/BBiLNSrcE730eWG9m1cBXiPRNiPQ4pkl2RERkD10p\niIhIK4WCiIi0UiiIiEgrhYKIiLTqdU80Dxo0yEeOHBnrMkREepVFixaVu3vu/rbrdaEwcuRIiouL\nY12GiEivYmYb9r+Vmo9ERKQNhYKIiLRSKIiISCuFgoiItFIoiIhIK4WCiIi0UiiIiEirqIaCmZ1p\nZquCSc5vaOf9K82szMyWBq8vRauWhet3cMuLKwmHNSqsiEhHohYKZhYiMi3hWURmobrUzCa2s+nD\n7j4teN0VrXqWbtzJba+sYVdjc7R2ISLS60XzSmEWUOLua929EXgIODeK++tUZmrk4e2aeoWCiEhH\nohkKBcCmNsulwbq9XWBmy8zsMTMb1t4Xmdk1ZlZsZsVlZWUHVUxGEAq7FAoiIh2KdUfzM8BId58C\n/B24t72N3P1Ody9y96Lc3P2O59SuzNQkAGrqmw6yVBGRvi+aobAZaPubf2GwrpW7V7h7Q7B4FzAz\nWsVkpATNRw26UhAR6Ug0Q2EhMM7MRplZMpHJzee23cDM8tsszgZWRKuYLPUpiIjsV9SGznb3ZjO7\nFngRCAFz3P19M7sJKHb3ucB1ZjYbaAZ2AFdGq549zUfqUxAR6VhU51Nw9+eB5/dad2ObP38f+H40\na9gjo/VKQX0KIiIdiXVHc7dJTw5hBrvUpyAi0qG4CQUzIyMlUX0KIiKdiJtQAMhKTVIoiIh0Iq5C\nIXKloD4FEZGOxFUoZKaq+UhEpDNxFQoZqYnqaBYR6URchUJmapKaj0REOhFnoaArBRGRzsRXKKQk\nUq0+BRGRDsVXKKQm0tgcpqG5JdaliIj0SHEVCntGStX4RyIi7YurUGgdFE/9CiIi7YqrUMjQ8Nki\nIp2Kq1DQPM0iIp2Lr1BI0ZScIiKdia9Q0JWCiEin4ioU9vQpqKNZRKR9cRUKmZp9TUSkU3EVCimJ\nIZITE6jRlYKISLviKhQgMtSF+hRERNoXf6GQmqgnmkVEOhB3oZCRqtnXREQ6EnehkJmSpLuPREQ6\nEHehkKEpOUVEOhR3oaB5mkVEOhZ/oZCiPgURkY7EXyikRvoU3D3WpYiI9DhxGAqJhB12N2r2NRGR\nvcVdKGhOBRGRjsVdKPx79jX1K4iI7C3+QiGYp7laVwoiIvuIv1DYM3y2QkFEZB9xFwrqUxAR6VhU\nQ8HMzjSzVWZWYmY3dLLdBWbmZlYUzXpAfQoiIp2JWiiYWQi4DTgLmAhcamYT29kuE/gmsCBatbSV\nkaIrBRGRjkTzSmEWUOLua929EXgIOLed7X4K/AKoj2ItrTLU0Swi0qFohkIBsKnNcmmwrpWZzQCG\nuftzUazjI0IJRkaK5lQQEWlPzDqazSwB+DXwnS5se42ZFZtZcVlZ2SHvO0PjH4mItCuaobAZGNZm\nuTBYt0cmMAl41czWA8cCc9vrbHb3O929yN2LcnNzD7mwzNREzakgItKOaIbCQmCcmY0ys2TgEmDu\nnjfdvcrdB7n7SHcfCbwFzHb34ijWBGhOBRGRjkQtFNy9GbgWeBFYATzi7u+b2U1mNjta++2KzNQk\nanSlICKyj8Rofrm7Pw88v9e6GzvY9pRo1tJWZkoipZW7u2t3IiK9Rtw90QxBn4Kaj0RE9hGXoRC5\n+0ihICKyt7gMhczUJOqaWmhuCce6FBGRHiVOQyEYKVWdzSIiHxGXoTAgPRmAspqGGFciItKzxGUo\njBucAcCqbTUxrkREpGeJy1AYm5dBKMFYuUWhICLSVlyGQkpiiDG56azcWh3rUkREepS4DAWACUOy\nWKErBRGRj4jbUBg/JJPNO+uo1mipIiKt4jYUjszPBGDVVl0tiIjsEbehMGFIFgArt6hfQURkj7gN\nhfzsVLJSE1mpKwURkVZxGwpmxoT8LIWCiEgbcRsKAEcOyWTV1hrCYY91KSIiPUJch8KE/Cx2NTSz\neWddrEsREekR4joUxg+J3IG0Qp3NIiJAvIfC4EgoqF9BRCQirkMhPSWREQPTNNyFiEggrkMBYMKQ\nTA2MJyISUCgMyWJdRS11jS2xLkVEJObiPhSOzM/CHZZvqYp1KSIiMRf3oXDMqAGYwbwPKmJdiohI\nzMV9KOSkJzO5IJt/fVAW61JERGIu7kMB4MRxg1iyaaeG0RaRuKdQAE4cl0tL2HlrjZqQRCS+KRSA\nGcNzSEsO8a8PymNdiohITCkUgOTEBI4dPVD9CiIS9xQKgRPHDWJ9xW427dgd61JERGJGoRA4cVwu\ngJqQRCSuKRQCY3LTyc9OVROSiMQ1hULAzDhx3CDml5TTokl3RCROKRTaOHFcLtX1zSzdVBnrUkRE\nYkKh0MbJ43NJSUzgqSUfxroUEZGYiGoomNmZZrbKzErM7IZ23v+Kmb1rZkvNbJ6ZTYxmPfuTlZrE\nmZOGMPedD6lv0qipIhJ/ohYKZhYCbgPOAiYCl7bzQ/8Bd5/s7tOAXwK/jlY9XXXBjEKq6pr454rt\nsS5FRKTbRfNKYRZQ4u5r3b0ReAg4t+0G7t52yrN0IOY9vB8bO4ghWak8vrg01qWIiHS7aIZCAbCp\nzXJpsO4jzOzrZraGyJXCde19kZldY2bFZlZcVhbdW0ZDCcb5Mwp4bXUZ26vro7ovEZGeJuYdze5+\nm7uPAb4H/LCDbe509yJ3L8rNzY16TRfMLKQl7Dy1dHPU9yUi0pNEMxQ2A8PaLBcG6zryEHBeFOvp\nsjG5GcwY3p/HFpXiHvMWLRGRbhPNUFgIjDOzUWaWDFwCzG27gZmNa7P4KeCDKNZzQC6YWcjqbbtY\nVqppOkUkfnQpFMxsjJmlBH8+xcyuM7P+nX3G3ZuBa4EXgRXAI+7+vpndZGazg82uNbP3zWwp8G3g\nCwd9JIfZOVOHkp4c4t431se6FBGRbmNdaR4JfmgXASOB54GngaPc/eyoVteOoqIiLy4u7pZ9/fcz\n73P/mxuY973TGJKd2i37FBGJBjNb5O5F+9uuq81H4eA3/88Af3D364H8QymwN/jix0YRduceXS2I\nSJzoaig0mdmlRJp3ng3WJUWnpJ5j2IA0zpqUzwMLNlDb0BzrckREoq6roXAVcBxws7uvM7NRwP3R\nK6vn+NKJo6iub+aR4k3731hEpJfrUii4+3J3v87dHzSzHCDT3X8R5dp6hOnDcygakcOc+es0pLaI\n9HldvfvoVTPLMrMBwGLgz2YW83GKusuXThzNph11PPfulliXIiISVV1tPsoOxik6H7jP3Y8BTo9e\nWT3LGRMHM2FIJr98YaVGTxWRPq2roZBoZvnARfy7ozluhBKMG8+ZSGllHXfPWxfrckREoqaroXAT\nkYfQ1rj7QjMbTQ96+rg7HD9mEJ88ajC3vVLCNg2UJyJ9VFc7mh919ynu/tVgea27XxDd0nqeH5w9\nkeYW5xcvrIx1KSIiUdHVjuZCM3vSzLYHr8fNrDDaxfU0wwemcfWJo3hi8WaWbNQ8ziLS93S1+egv\nRAazGxq8ngnWxZ2vnzqWwVkpfPfRd6hrVKeziPQtXQ2FXHf/i7s3B697gOhPbNADZaQk8uuLprG2\nvJafPrc81uWIiBxWXQ2FCjO73MxCwetyoCKahfVkHxs7iGtOHM0DCzby4vtbY12OiMhh09VQ+CKR\n21G3AluAC4Ero1RTr/CdT4xnUkEW33t8GVurdDeSiPQNXb37aIO7z3b3XHfPc/fzgLi7+6it5MQE\nfnfJdBqawtz49HuxLkdE5LA4lJnXvn3YquilxuRmcO1pY3lp+TbeKCmPdTkiIofsUELBDlsVvdjV\nJ4yiMKcfNz27nOaWcKzLERE5JIcSChoyFEhNCvGDs49k5dYaHlqo4bVFpHfrNBTMrMbMqtt51RB5\nXkGAMycN4ZhRA7j1pVVU7W6KdTkiIget01Bw90x3z2rnlenuid1VZE9nFhkwb2ddEzc/v5yuzHst\nItITHUrzkbRx1NBsvnbKGB4pLuVPr62NdTkiIgdFv+0fRt85Yzwbd9TxixdWMrR/KudOK4h1SSIi\nB0ShcBglJBi/+uwUtlXXc/2jyxiclcqxowfGuiwRkS5T89FhlpIY4s7Pz2TYgH58+f5FrC3bFeuS\nRES6TKEQBf3TkvnLlbMIJRhX31vMzt2NsS5JRKRLFApRMnxgGnd8fiabK+v4yv8uorFZD7aJSM+n\nUIiio0cO4BcXTuattTv4wZPv6lZVEenx1NEcZZ+ZXsi6slp+/3IJhTlpfPP0cbEuSUSkQwqFbvCt\nM46gdGcdv/nHavL7p3JR0bBYlyQi0i6FQjcwM35+/hS2VzfwX0+8y+CsVE4+Ii4nrhORHk59Ct0k\nOTGB2y+fwdi8DK59YDGbduyOdUkiIvtQKHSjzNQk/nxFETh848ElNGmobRHpYRQK3WzYgDR+fsEU\nlm7aya0vrY51OSIiHxHVUDCzM81slZmVmNkN7bz/bTNbbmbLzOyfZjYimvX0FJ+aks/njhnOn15b\nw+ury2JdjohIq6iFgpmFgNuAs4CJwKVmNnGvzZYARe4+BXgM+GW06ulpbvz0RMYPzuTaBxbz0vtb\nY12OiAgQ3SuFWUCJu69190bgIeDcthu4+yvuvqfH9S2gMIr19CipSSHu+kIRIwamc839i7jpmeV6\n6llEYi6aoVAAtJ2fsjRY15Grgb+194aZXWNmxWZWXFbWd5pbhg1I47GvHseVx49kzvx1fPaON9le\nUx/rskQkjvWIjmYzuxwoAm5p7313v9Pdi9y9KDe3b93fn5IY4iezj+JPl89g9dYaPnPbG3ywrSbW\nZYlInIpmKGwG2j66Wxis+wgzOx34ATDb3RuiWE+PduakfB7+8rE0NIe54PY3eHNNRaxLEpE4FM1Q\nWAiMM7NRZpYMXALMbbuBmU0H7iASCNujWEuvMKWwP09+7XjyslL5/N0LuP3VNbSENYieiHSfqIWC\nuzcD1wIvAiuAR9z9fTO7ycxmB5vdAmQAj5rZUjOb28HXxY1hA9J4/CvH84mjBvOLF1Zy2V1v8eHO\nuliXJSJxwnrbcM5FRUVeXFwc6zKizt15dFEpP5n7PkmhBP76pWOYVJAd67JEpJcys0XuXrS/7XpE\nR7Psy8y4qGgYz113IunJIa78y0I2VNTGuiwR6eMUCj3cqEHp3Hf1LJrDYa6Y8zZlNXHbFy8i3UCh\n0AuMzctkzpVHs626nqvueZvKWs35LCLRoVDoJWYMz+H2y2ayeusuzvnjPN7bXBXrkkSkD1Io9CKn\nTsjjka8cRzjsXHD7GzxavGn/HxIROQAKhV5m2rD+PPONE5g5IofrH1vGn15bE+uSRKQPUSj0QgMz\nUrjvi7M4Z+pQfv63lTyyUFcMInJ4aI7mXioxlMCtn51KVV0TNzyxjOy0JD551JBYlyUivZyuFHqx\n5MQE/nT5DKYU9ucbDy7hR0+9x/PvbmGH7k4SkYOkJ5r7gMraRm54Yhn/+qCc3Y0tmMElRw/jR5+e\nSFqyLgZFpOtPNOsnRh+Qk57MHZ8voqklzLLSKp5550PufXM9C9bu4LeXTGNKYf9YlygivYSaj/qQ\npFACM0fk8JPZR/HAl46lrqmF8//nDebMW0dvuyIUkdhQKPRRx40ZyAvfPIlTJ+Rx07PLufHp92lu\n0XSfItI5hUIflp2WxB2Xz+TLJ43m/rc2cPW9xVTXN8W6LBHpwRQKfVxCgvH9s4/kZ+dPZl5JOR/7\n+cvc9Mxy1pdrxFUR2ZdCIU5cOms4T33tY5w6Po/73lzPqbe+yo+eeo+wZnYTkTYUCnFkcmE2v790\nOm/ccBpXHDuC+9/awA1PLFMwiEgr3ZIah/KyUvnJ7KPI7pfE718uwTB+dv5kEhIs1qWJSIwpFOKU\nmfGtM47AgT+8XMLb63cQSjCaWsJMKezPLRdOITUpFOsyRaSbKRTimJnx7TOOYEB6MvNLyklOTMAd\nnl32IZW1jfz5iiL6JSsYROKJhrmQfTy2qJTrH3uH48cM5K4rjlYwiPQBXR3mQh3Nso8LZxbyqwun\n8saaCi7581u8snK7OqNF4oSaj6RdF8wsJCUpgZ8+u5yr7lnI6Nx0vnDcSM6enE9uZkqsyxORKFHz\nkXSqqSXM8+9u4e5561hWWkWCwaxRA/jUlKGcN20omalJsS5RRLqgq81HCgXpEndn1bYann93K8+/\nu4WS7btITw7xmRkFfP7YkYwfkhnrEkWkEwoFiaplpTu5780NzH3nQxqbw5wzdSjXf2I8wwemxbo0\nEWmHQkG6RWVtI3fPW8dd89bSEnYuO2YEXztlDHlZqbEuTUTaUChIt9pWXc9v/7GahxduIjGUwCVH\nD+MrJ49haP9+sS5NRFAoSIxsqKjl9lfX8PjiUsIOUwuzOW7MQI4fM4hjRg0gMaS7oEViQaEgMbV5\nZx0PLtjI/DXlLCutoiXsDM5K4aKiYVx89DAKc9T3INKdFArSY+xqaGbeB+U8vHAjr64uA+D4MQM5\nb1oBZ03OJyNFj8uIRJtCQXqk0srdPFpcypNLNrNxx25SkxK47JgRXPfxcWT30zMPItGiUJAezd1Z\nvLGSBxZs4oklpeSkJfPdT4znmNED2Lm7kcraJsbmZTByUHqsSxXpE3pEKJjZmcDvgBBwl7v/fK/3\nTwJ+C0wBLnH3x/b3nQqFvue9zVXc9Mxy3l6/Y5/3Tj4ilyuPH8nJR+RqvgeRQxDzUDCzELAaOAMo\nBRYCl7r78jbbjASygO8CcxUK8cvdeW11GVV1TfRPSyYrNZF/fVDO/761ge01DYzOTeerJ4/hvOkF\nJOkOJpED1tVQiGYP3yygxN3XBgU9BJwLtIaCu68P3gtHsQ7pBcyMU8bnfWTd9OE5fPWUMfztva38\n6dU1XP/YMn7z99WcdEQuNfXN7KxrZPiAdP7r7Akag0nkMInmr1wFwKY2y6XBugNmZteYWbGZFZeV\nlR2W4qR3SAolMHvqUJ677gT+ctXRFA5I4x8rtrFyazW7G1t4pHgT5/5xPqu31cS6VJE+oVfcC+ju\ndwJ3QqT5KMblSAyYGaeOz+PUva4mFqyt4OsPLOHcP87n22ccwdD+/eiXnEBOWjITh2aRkqgJgkQO\nRDRDYTMwrM1yYbBO5LA5ZvRAnrvuBK59YDE3P7/iI++lJCYwY3gOJx4xiKuOH6UZ5ES6IJqhsBAY\nZ2ajiITBJcDnorg/iVODs1J5+Jrj2Lyzjt2NLdQ3tbClqo6311WyYF0Fv3xhFQ8v3MTPz5/CcWMG\nUl3fxGPFpcwvKeczMwr41OR8zHRnkwhE/5bUs4ncchoC5rj7zWZ2E1Ds7nPN7GjgSSAHqAe2uvtR\nnX2n7j6SA/Xmmgq+9/gyNu7YzYnjBrFoQyW7G1sYmJ5MRW0jU4f15wdnH8msUQNiXapI1MT8ltRo\nUSjIwdjd2MytL63m6aWbOW1CHlccN5Ij87N4YnEpv3ppFduqGxiXl8FZk4bwyUlDOHJIlp6LkD5F\noSDSRXWNLTy6aBPPLdvCwvU7CDtkpSYydVh/pg3rz/kzChmlJ6ull1MoiByE8l0NvLxyO0s2VrJ0\nU1Xrra4XHz2MbwbjMy3eUMmiDZVMyM/i9CPz1B8hvYJCQeQwKKtp4LZXSvjrgg0kmOEOjS3/ftby\nhLGD+PE5Exk3WHNUS8+mUBA5jDZW7ObueWtJTQpx7OiBTB/en6eXfsitL62itrGFiflZNLWEaWwO\nM2xAGpfOGs7Hj8wjKZRAc0uYlVtrqNzdyPjBmeRmpujqQrqdQkGkG+yobeQPL3/A2rJaUhITSEpM\nYPGGSrZU1ZOXmcLYvAze2bST2saW1s/kpCVRNHIA3ztzAmPzMmJYvcQThYJIjDS3hHl1VRkPvr2R\nbTX1zBiew8wROeRmpLB6Ww0rt9bw/LtbqG8K85WTR/Plk8ewelsN80vKKa2s4/JjRzCpIDvWhyF9\njEJBpAcrq2ng5ueW89TSD0kwCAf/G/ZLClHX1MI5U4dy7aljaWwOs7Z8FxW7Gvn01HzyMlNjW7j0\nWgoFkV5gfkk5r6zczrTh/Tlu9ECSEhO487W13DVvLfVNHx08ODM1kf/85Hg+d8wIQns9Q1Fd38Si\n9ZVMyM8kP7tfdx6C9BIKBZFebHt1PS+8v5W8zBRGDcrAcX767HLml1QwuSCbmSNySE0KkWBQHNwi\n2xJ2kkMJfO6Y4XztlDHkZemqQv5NoSDSx7g7c9/5kN/98wPKaxqob47c7TQxP4tTJ+Qya9RA/vbu\nFh5dVEpigjFqUDqNzWEamsOMH5LJ52YN59QJeYQSjLrGFt4p3UlGSqL6L+KEQkEkDoTDvs9wHBsq\narnz9bWU1TSQkhQiMcGYX1LO9poG8rNTyctK5f3NVTQHHRmnTcjj22ccwaSCbKrrm3hvcxXbqxvI\n7pdEVr8kCnP6MVhXHb2eQkFEWjW1hPnniu08vHAjtY0tFI3IoWhkDiu21HDHa2uorm+moH8/Nu+s\n2+ezCQYXFQ3jP04/giHZqXywrYbfv1zCm2vKufL4kVxz0hiSEzVFak+nUBCRLqmqa2LOvHV8sL2G\niflZTCrIpjAnjZr6Jqrqmnh9dTn3v7WeUIJRNGIA89eUk5YUYlJBNgvW7WD84Ex++OkjqW1oYcmm\nSlZtrWFcXgbHjxnE0aMGkJHSK+by6vMUCiJy2GzasZtfvbSKt9ZW8NmZw7j6hFHkpCfz9+XbuPHp\n99hSVQ9AUijSl7G+fDeNLWFCCca4vAymFGYzqSCbXQ3NrCurZUPFbrL6JTJucCbj8jLYUdvI0k07\nWVZaxcT8LH52/mRy0pNjfNR9i0JBRLrFroZm/r58K8MHpHPU0CxSk0LUN7WweEMlb62t4J3SKpaV\n7qRydxMAeZkpjBiYxs7dTawrr23t2xiancrEoVm8vrqcQRnJ3HbZDCYXZPPC+1uZM28dNfXNXH7s\nCC6cWUi6rj4OmEJBRHoMd2drdT2ZqUkfaU5qagmzoaKWzNSk1s7sZaU7+dpfF7Otup68zFQ276xj\n5MA0+qcls3TTTrJSEzl7cj5DslMZlJFC/7QkEhOMUEICDc0trCurZW15LTtqGzl6ZA6njM/jqKFZ\nhzzeVFNLmNtfXcOoQemcM3XoIX1XLCgURKTXqtrdxI/nvkfZrgauOG4kpx85mFCCsWhDJXPmrWP+\nmnJ2Blce7RmanUpmahKrgqHPs1ITSUgw6ptaaGgOk5hgJIUSSElMYNiANMbmZjB2cAanTchjwpCs\nfb5vR20jX//rYt5cW0Eowbj3qlmcMG7QYTnWku27uOO1NXz1lDGMzo3eWFgKBRHp0xqbw+yobaSq\nromWsBN2JzFkjBiQTr/kEADba+p5fXU5izdWkphgpCaFSElMoCXsNLWEqWtqYUPFbj7Ytout1ZF+\nkckF2Xy2qJBxeZkkhYyahmZ++GQkoH58zkTue2MDW6rqeOrrH2v9If5uaRXbquuZMSKHAQfQF1JZ\n28h5/zOfDRW7yUxJ5JbPTuXMSUMO/18WCgURkQNSsauBue98yKPFpSzfUv2R9/KzU7nj8zOZUtif\nTTt2M/uP88hJS+b6T47nnjfWs2DdjtZtR+emc2R+Fv37JZHdL4lQgrG1qp6t1fUkJhjf+Pg4ZgzP\nobE5zBVzFrB4w05+c/E07nx9De+UVvHlk0bzjY+PO+x3bSkUREQOUsn2GrbXNNDc4rSEnenD+9M/\n7d9XAAvWVnD53QtoanHys1O5+oRRTCrIZvHGSorXV7K2bBfV9c1U1TURdicvM4UhWal8WFVPWU0D\n508vwMx4fHEpv7l4Kp+ZXkhDcwv//cxyHliwkbTkEJ+anM+FMws5cmgWmSmJh9wnolAQEYmi11eX\nUbm7kbMn55MUav/hPXcn7LQOYFjb0Mz/vFrCn/+1jsbmMF89ZQzfO3PCRz6zaEMljyzcxLPLPmyd\nh6NfUojBWSl864wjOHdawUHVq1AQEemhNlbs5u31Ozh/esE+w5TsUdvQzGury9hcWce26nq21TRw\ncdGwg+7g7moo6GZfEZFuNnxgGsMHpnW6TXpK5Nbb7qYBS0REpJVCQUREWikURESklUJBRERaKRRE\nRKSVQkFERFopFEREpJVCQUREWvW6J5rNrAzYcJAfHwSUH8Zyeot4PO54PGaIz+OOx2OGAz/uEe6e\nu7+Nel0oHAozK+7KY959TTwedzweM8TnccfjMUP0jlvNRyIi0kqhICIireItFO6MdQExEo/HHY/H\nDPF53PF4zBCl446rPgUREelcvF0piIhIJxQKIiLSKm5CwczONLNVZlZiZjfEup5oMLNhZvaKmS03\ns/fN7JvB+gFm9ncz+yD4b07JYuFVAAAGTUlEQVSsaz3czCxkZkvM7NlgeZSZLQjO98Nmlry/7+ht\nzKy/mT1mZivNbIWZHRcn5/pbwb/v98zsQTNL7Wvn28zmmNl2M3uvzbp2z61F/D449mVmNuNQ9h0X\noWBmIeA24CxgInCpmU2MbVVR0Qx8x90nAscCXw+O8wbgn+4+DvhnsNzXfBNY0Wb5F8Bv3H0sUAlc\nHZOqout3wAvuPgGYSuT4+/S5NrMC4DqgyN0nASHgEvre+b4HOHOvdR2d27OAccHrGuD2Q9lxXIQC\nMAsocfe17t4IPAScG+OaDjt33+Lui4M/1xD5IVFA5FjvDTa7FzgvNhVGh5kVAp8C7gqWDTgNeCzY\npC8eczZwEnA3gLs3uvtO+vi5DiQC/cwsEUgDttDHzre7vw7s2Gt1R+f2XOA+j3gL6G9mBz2PZ7yE\nQgGwqc1yabCuzzKzkcB0YAEw2N23BG9tBQbHqKxo+S3wn0A4WB4I7HT35mC5L57vUUAZ8Jeg2ewu\nM0unj59rd98M/ArYSCQMqoBF9P3zDR2f28P68y1eQiGumFkG8DjwH+5e3fY9j9yD3GfuQzazTwPb\n3X1RrGvpZonADOB2d58O1LJXU1FfO9cAQTv6uURCcSiQzr7NLH1eNM9tvITCZmBYm+XCYF2fY2ZJ\nRALhr+7+RLB6257LyeC/22NVXxR8DJhtZuuJNAueRqStvX/QvAB983yXAqXuviBYfoxISPTlcw1w\nOrDO3cvcvQl4gsi/gb5+vqHjc3tYf77FSygsBMYFdygkE+mYmhvjmg67oC39bmCFu/+6zVtzgS8E\nf/4C8HR31xYt7v59dy9095FEzuvL7n4Z8ApwYbBZnzpmAHffCmwys/HBqo8Dy+nD5zqwETjWzNKC\nf+97jrtPn+9AR+d2LnBFcBfSsUBVm2amAxY3TzSb2dlE2p5DwBx3vznGJR12ZnYC8C/gXf7dvv5f\nRPoVHgGGExl2/CJ337sTq9czs1OA77r7p81sNJErhwHAEuByd2+IZX2Hm5lNI9K5ngysBa4i8ote\nnz7XZvbfwMVE7rZbAnyJSBt6nznfZvYgcAqR4bG3AT8GnqKdcxuE4x+JNKPtBq5y9+KD3ne8hIKI\niOxfvDQfiYhIFygURESklUJBRERaKRRERKSVQkFERFopFKRPMDM3s1vbLH/XzH4Shf3cEozQecvh\n/u797PceM7tw/1uKHJrE/W8i0is0AOeb2c/cvTyK+7kGGODuLVHch0jM6EpB+opmInPWfmvvN8xs\npJm9HIw1/08zG97ZFwVPht4SjNf/rpldHKyfC2QAi/asa/OZ9GAM/LeDAerODdZfaWZPm9mrwTj4\nP27zmW8H+3jPzP6jzforglrfMbP72+zmJDN7w8zW7rlqMLN8M3vdzJYG33PiAf/NibShKwXpS24D\nlpnZL/da/wfgXne/18y+CPyezodWPh+YRmSOgkHAQjN73d1nm9kud5/Wzmd+QGSIjS+aWX/gbTP7\nR/DeLGASkadNF5rZc0QGM7sKOAYwYIGZvQY0Aj8Ejnf3cjMb0GYf+cAJwAQiQxs8BnwOeNHdbw7m\nDUnb79+SSCcUCtJnuHu1md1HZBKWujZvHUfkBz3A/cDeobG3E4AHgyaibcEP66PpfLysTxAZmO+7\nwXIqkeEIAP7u7hUAZvZE8P0OPOnutW3Wnxisf3RPE9heQ1Q85e5hYLmZ7Rk2eSEwJxgI8Sl3X7qf\nYxPplJqPpK/5LZFZt9K7eb8GXODu04LXcHffMxPc3mPJHOzYMm3H8jFonYzlJCKjYt5jZlcc5HeL\nAAoF6WOC36wf4aPTMb5BZARVgMuIDBrYmX8BF1tk3udcIj90397PZ14EvhEMToaZTW/z3hkWmV+3\nH5Fmq/nBPs4LRvtMBz4TrHsZ+KyZDQy+p23z0T7MbASwzd3/TGRwvEOan1dEzUfSF90KXNtm+RtE\nZii7nshsZVcBmNlsInP93rjX558k0uT0DpHf6v8zGKq6Mz8lcpWyzMwSgHXAp4P33iYyx0Uh8L97\nRrA0s3v4d9jc5e5LgvU3A6+ZWQuRET+v7GS/pwDXm1kTsAvQlYIcEo2SKhJFZnYlkeC5dn/bivQE\naj4SEZFWulIQEZFWulIQEZFWCgUREWmlUBARkVYKBRERaaVQEBGRVv8H+HIoYx1qo8UAAAAASUVO\nRK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEWCAYAAABMoxE0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nIiLSLkokIiLSLkokIiLSLkokIiLSLv8f8jaYb4a+ps8AAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nmycPUFLNJZA",
        "colab_type": "code",
        "outputId": "0ebab946-7d5e-43ea-b94c-0753c956d8dc",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "test_loss, test_acc = model.evaluate(testX, testY)\n",
        "test_loss , test_acc"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "10000/10000 [==============================] - 0s 35us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.7063835167914629, 0.8889)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XY4Zn52QOsIa",
        "colab_type": "code",
        "outputId": "66552006-4e89-41ef-e6b1-eebe0b25946b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 357
        }
      },
      "source": [
        "model_1 = keras.Sequential([\n",
        "    keras.layers.Flatten(input_shape=(trainX.shape[1],trainX.shape[2])),\n",
        "    keras.layers.Dense(512, activation=tf.nn.relu),\n",
        "    keras.layers.Dense(256, activation=tf.nn.relu),\n",
        "    keras.layers.Dense(128, activation=tf.nn.relu),\n",
        "    keras.layers.Dense(64, activation=tf.nn.relu),\n",
        "    keras.layers.Dense(10, activation=tf.nn.softmax)\n",
        "])\n",
        "model_1.compile(optimizer='adam',\n",
        "              loss='sparse_categorical_crossentropy',\n",
        "              metrics=['accuracy'])\n",
        "model_1.summary()"
      ],
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "flatten_2 (Flatten)          (None, 784)               0         \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 512)               401920    \n",
            "_________________________________________________________________\n",
            "dense_4 (Dense)              (None, 256)               131328    \n",
            "_________________________________________________________________\n",
            "dense_5 (Dense)              (None, 128)               32896     \n",
            "_________________________________________________________________\n",
            "dense_6 (Dense)              (None, 64)                8256      \n",
            "_________________________________________________________________\n",
            "dense_7 (Dense)              (None, 10)                650       \n",
            "=================================================================\n",
            "Total params: 575,050\n",
            "Trainable params: 575,050\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c2aZ1iH2PGnR",
        "colab_type": "code",
        "outputId": "641032a8-25b3-48a8-a8f1-f515ca54406b",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "history_1 = model_1.fit(trainX, trainY, epochs=epochs, validation_split=validation_split,verbose=verbose, shuffle = True)"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/100\n",
            "48000/48000 [==============================] - 15s 315us/step - loss: 0.5061 - acc: 0.8162 - val_loss: 0.4123 - val_acc: 0.8504\n",
            "Epoch 2/100\n",
            "48000/48000 [==============================] - 15s 315us/step - loss: 0.3786 - acc: 0.8624 - val_loss: 0.3736 - val_acc: 0.8665\n",
            "Epoch 3/100\n",
            "48000/48000 [==============================] - 15s 315us/step - loss: 0.3415 - acc: 0.8744 - val_loss: 0.3396 - val_acc: 0.8764\n",
            "Epoch 4/100\n",
            "48000/48000 [==============================] - 15s 319us/step - loss: 0.3180 - acc: 0.8830 - val_loss: 0.3211 - val_acc: 0.8869\n",
            "Epoch 5/100\n",
            "48000/48000 [==============================] - 15s 323us/step - loss: 0.2962 - acc: 0.8900 - val_loss: 0.3394 - val_acc: 0.8751\n",
            "Epoch 6/100\n",
            "48000/48000 [==============================] - 15s 317us/step - loss: 0.2839 - acc: 0.8938 - val_loss: 0.3188 - val_acc: 0.8868\n",
            "Epoch 7/100\n",
            "48000/48000 [==============================] - 16s 327us/step - loss: 0.2684 - acc: 0.9002 - val_loss: 0.3379 - val_acc: 0.8812\n",
            "Epoch 8/100\n",
            "48000/48000 [==============================] - 15s 318us/step - loss: 0.2564 - acc: 0.9049 - val_loss: 0.3314 - val_acc: 0.8833\n",
            "Epoch 9/100\n",
            "48000/48000 [==============================] - 15s 318us/step - loss: 0.2451 - acc: 0.9067 - val_loss: 0.3569 - val_acc: 0.8814\n",
            "Epoch 10/100\n",
            "48000/48000 [==============================] - 16s 335us/step - loss: 0.2408 - acc: 0.9098 - val_loss: 0.3186 - val_acc: 0.8852\n",
            "Epoch 11/100\n",
            "48000/48000 [==============================] - 16s 336us/step - loss: 0.2292 - acc: 0.9147 - val_loss: 0.3233 - val_acc: 0.8862\n",
            "Epoch 12/100\n",
            "48000/48000 [==============================] - 16s 331us/step - loss: 0.2195 - acc: 0.9169 - val_loss: 0.3440 - val_acc: 0.8844\n",
            "Epoch 13/100\n",
            "48000/48000 [==============================] - 16s 339us/step - loss: 0.2139 - acc: 0.9200 - val_loss: 0.3064 - val_acc: 0.8965\n",
            "Epoch 14/100\n",
            "48000/48000 [==============================] - 16s 341us/step - loss: 0.2015 - acc: 0.9236 - val_loss: 0.3276 - val_acc: 0.8900\n",
            "Epoch 15/100\n",
            "48000/48000 [==============================] - 16s 332us/step - loss: 0.1983 - acc: 0.9240 - val_loss: 0.3679 - val_acc: 0.8806\n",
            "Epoch 16/100\n",
            "48000/48000 [==============================] - 16s 336us/step - loss: 0.1912 - acc: 0.9279 - val_loss: 0.3375 - val_acc: 0.8920\n",
            "Epoch 17/100\n",
            "48000/48000 [==============================] - 16s 330us/step - loss: 0.1848 - acc: 0.9295 - val_loss: 0.3722 - val_acc: 0.8859\n",
            "Epoch 18/100\n",
            "48000/48000 [==============================] - 16s 326us/step - loss: 0.1808 - acc: 0.9318 - val_loss: 0.3375 - val_acc: 0.8949\n",
            "Epoch 19/100\n",
            "48000/48000 [==============================] - 16s 336us/step - loss: 0.1747 - acc: 0.9338 - val_loss: 0.3483 - val_acc: 0.8933\n",
            "Epoch 20/100\n",
            "48000/48000 [==============================] - 16s 337us/step - loss: 0.1747 - acc: 0.9337 - val_loss: 0.3601 - val_acc: 0.8962\n",
            "Epoch 21/100\n",
            "48000/48000 [==============================] - 16s 327us/step - loss: 0.1640 - acc: 0.9383 - val_loss: 0.3963 - val_acc: 0.8892\n",
            "Epoch 22/100\n",
            "48000/48000 [==============================] - 16s 329us/step - loss: 0.1590 - acc: 0.9394 - val_loss: 0.3681 - val_acc: 0.8916\n",
            "Epoch 23/100\n",
            "48000/48000 [==============================] - 16s 341us/step - loss: 0.1567 - acc: 0.9411 - val_loss: 0.3493 - val_acc: 0.8916\n",
            "Epoch 24/100\n",
            "48000/48000 [==============================] - 16s 343us/step - loss: 0.1520 - acc: 0.9419 - val_loss: 0.3791 - val_acc: 0.8940\n",
            "Epoch 25/100\n",
            "48000/48000 [==============================] - 16s 343us/step - loss: 0.1520 - acc: 0.9424 - val_loss: 0.3956 - val_acc: 0.8944\n",
            "Epoch 26/100\n",
            "48000/48000 [==============================] - 16s 336us/step - loss: 0.1451 - acc: 0.9439 - val_loss: 0.4679 - val_acc: 0.8838\n",
            "Epoch 27/100\n",
            "48000/48000 [==============================] - 16s 333us/step - loss: 0.1421 - acc: 0.9458 - val_loss: 0.4301 - val_acc: 0.8968\n",
            "Epoch 28/100\n",
            "48000/48000 [==============================] - 17s 355us/step - loss: 0.1404 - acc: 0.9463 - val_loss: 0.3876 - val_acc: 0.8989\n",
            "Epoch 29/100\n",
            "48000/48000 [==============================] - 17s 356us/step - loss: 0.1395 - acc: 0.9470 - val_loss: 0.3814 - val_acc: 0.8982\n",
            "Epoch 30/100\n",
            "48000/48000 [==============================] - 17s 346us/step - loss: 0.1362 - acc: 0.9476 - val_loss: 0.4242 - val_acc: 0.8937\n",
            "Epoch 31/100\n",
            "48000/48000 [==============================] - 16s 343us/step - loss: 0.1273 - acc: 0.9515 - val_loss: 0.4420 - val_acc: 0.8924\n",
            "Epoch 32/100\n",
            "48000/48000 [==============================] - 16s 336us/step - loss: 0.1242 - acc: 0.9513 - val_loss: 0.4401 - val_acc: 0.8929\n",
            "Epoch 33/100\n",
            "48000/48000 [==============================] - 16s 335us/step - loss: 0.1240 - acc: 0.9526 - val_loss: 0.4296 - val_acc: 0.8908\n",
            "Epoch 34/100\n",
            "48000/48000 [==============================] - 16s 340us/step - loss: 0.1238 - acc: 0.9522 - val_loss: 0.4337 - val_acc: 0.8947\n",
            "Epoch 35/100\n",
            "48000/48000 [==============================] - 17s 350us/step - loss: 0.1252 - acc: 0.9532 - val_loss: 0.4575 - val_acc: 0.8920\n",
            "Epoch 36/100\n",
            "48000/48000 [==============================] - 16s 334us/step - loss: 0.1160 - acc: 0.9549 - val_loss: 0.4510 - val_acc: 0.8885\n",
            "Epoch 37/100\n",
            "48000/48000 [==============================] - 16s 337us/step - loss: 0.1157 - acc: 0.9558 - val_loss: 0.5164 - val_acc: 0.8897\n",
            "Epoch 38/100\n",
            "48000/48000 [==============================] - 16s 338us/step - loss: 0.1182 - acc: 0.9553 - val_loss: 0.4495 - val_acc: 0.8955\n",
            "Epoch 39/100\n",
            "48000/48000 [==============================] - 17s 345us/step - loss: 0.1060 - acc: 0.9597 - val_loss: 0.4745 - val_acc: 0.8874\n",
            "Epoch 40/100\n",
            "48000/48000 [==============================] - 16s 341us/step - loss: 0.1094 - acc: 0.9576 - val_loss: 0.4457 - val_acc: 0.8910\n",
            "Epoch 41/100\n",
            "48000/48000 [==============================] - 18s 378us/step - loss: 0.1077 - acc: 0.9591 - val_loss: 0.5511 - val_acc: 0.8818\n",
            "Epoch 42/100\n",
            "48000/48000 [==============================] - 18s 375us/step - loss: 0.1096 - acc: 0.9588 - val_loss: 0.4703 - val_acc: 0.8935\n",
            "Epoch 43/100\n",
            "48000/48000 [==============================] - 16s 341us/step - loss: 0.1024 - acc: 0.9610 - val_loss: 0.4697 - val_acc: 0.8960\n",
            "Epoch 44/100\n",
            "48000/48000 [==============================] - 15s 323us/step - loss: 0.0995 - acc: 0.9624 - val_loss: 0.4857 - val_acc: 0.8976\n",
            "Epoch 45/100\n",
            "48000/48000 [==============================] - 15s 322us/step - loss: 0.1026 - acc: 0.9607 - val_loss: 0.5010 - val_acc: 0.8942\n",
            "Epoch 46/100\n",
            "48000/48000 [==============================] - 16s 331us/step - loss: 0.0951 - acc: 0.9642 - val_loss: 0.4921 - val_acc: 0.8922\n",
            "Epoch 47/100\n",
            "48000/48000 [==============================] - 16s 326us/step - loss: 0.0992 - acc: 0.9629 - val_loss: 0.5417 - val_acc: 0.8892\n",
            "Epoch 48/100\n",
            "48000/48000 [==============================] - 16s 339us/step - loss: 0.1019 - acc: 0.9621 - val_loss: 0.4838 - val_acc: 0.8936\n",
            "Epoch 49/100\n",
            "48000/48000 [==============================] - 16s 335us/step - loss: 0.0963 - acc: 0.9643 - val_loss: 0.5211 - val_acc: 0.8926\n",
            "Epoch 50/100\n",
            "48000/48000 [==============================] - 16s 342us/step - loss: 0.0909 - acc: 0.9659 - val_loss: 0.5270 - val_acc: 0.8918\n",
            "Epoch 51/100\n",
            "48000/48000 [==============================] - 17s 355us/step - loss: 0.0981 - acc: 0.9643 - val_loss: 0.4934 - val_acc: 0.8892\n",
            "Epoch 52/100\n",
            "48000/48000 [==============================] - 16s 335us/step - loss: 0.0849 - acc: 0.9687 - val_loss: 0.5418 - val_acc: 0.8951\n",
            "Epoch 53/100\n",
            "48000/48000 [==============================] - 17s 350us/step - loss: 0.0854 - acc: 0.9675 - val_loss: 0.5027 - val_acc: 0.8912\n",
            "Epoch 54/100\n",
            "48000/48000 [==============================] - 16s 334us/step - loss: 0.0875 - acc: 0.9671 - val_loss: 0.5624 - val_acc: 0.8912\n",
            "Epoch 55/100\n",
            "48000/48000 [==============================] - 16s 333us/step - loss: 0.0888 - acc: 0.9669 - val_loss: 0.5143 - val_acc: 0.8955\n",
            "Epoch 56/100\n",
            "48000/48000 [==============================] - 16s 339us/step - loss: 0.0882 - acc: 0.9672 - val_loss: 0.5744 - val_acc: 0.8878\n",
            "Epoch 57/100\n",
            "48000/48000 [==============================] - 17s 350us/step - loss: 0.0800 - acc: 0.9703 - val_loss: 0.5508 - val_acc: 0.8948\n",
            "Epoch 58/100\n",
            "48000/48000 [==============================] - 16s 334us/step - loss: 0.0843 - acc: 0.9686 - val_loss: 0.5452 - val_acc: 0.8936\n",
            "Epoch 59/100\n",
            "48000/48000 [==============================] - 15s 320us/step - loss: 0.0821 - acc: 0.9693 - val_loss: 0.5431 - val_acc: 0.8930\n",
            "Epoch 60/100\n",
            "48000/48000 [==============================] - 15s 320us/step - loss: 0.0726 - acc: 0.9730 - val_loss: 0.5906 - val_acc: 0.8989\n",
            "Epoch 61/100\n",
            "48000/48000 [==============================] - 16s 329us/step - loss: 0.0796 - acc: 0.9710 - val_loss: 0.5682 - val_acc: 0.8958\n",
            "Epoch 62/100\n",
            "48000/48000 [==============================] - 16s 324us/step - loss: 0.0834 - acc: 0.9700 - val_loss: 0.6140 - val_acc: 0.8932\n",
            "Epoch 63/100\n",
            "48000/48000 [==============================] - 16s 334us/step - loss: 0.0795 - acc: 0.9715 - val_loss: 0.5453 - val_acc: 0.8942\n",
            "Epoch 64/100\n",
            "48000/48000 [==============================] - 16s 332us/step - loss: 0.0736 - acc: 0.9729 - val_loss: 0.5669 - val_acc: 0.8938\n",
            "Epoch 65/100\n",
            "48000/48000 [==============================] - 16s 332us/step - loss: 0.0788 - acc: 0.9715 - val_loss: 0.5606 - val_acc: 0.8935\n",
            "Epoch 66/100\n",
            "48000/48000 [==============================] - 15s 322us/step - loss: 0.0681 - acc: 0.9742 - val_loss: 0.5884 - val_acc: 0.8933\n",
            "Epoch 67/100\n",
            "48000/48000 [==============================] - 15s 318us/step - loss: 0.0694 - acc: 0.9741 - val_loss: 0.5907 - val_acc: 0.8885\n",
            "Epoch 68/100\n",
            "48000/48000 [==============================] - 16s 329us/step - loss: 0.0846 - acc: 0.9716 - val_loss: 0.5901 - val_acc: 0.8918\n",
            "Epoch 69/100\n",
            "48000/48000 [==============================] - 16s 338us/step - loss: 0.0661 - acc: 0.9757 - val_loss: 0.6261 - val_acc: 0.8958\n",
            "Epoch 70/100\n",
            "48000/48000 [==============================] - 17s 346us/step - loss: 0.0686 - acc: 0.9753 - val_loss: 0.6185 - val_acc: 0.8952\n",
            "Epoch 71/100\n",
            "48000/48000 [==============================] - 16s 333us/step - loss: 0.0779 - acc: 0.9732 - val_loss: 0.6107 - val_acc: 0.8988\n",
            "Epoch 72/100\n",
            "48000/48000 [==============================] - 16s 343us/step - loss: 0.0623 - acc: 0.9774 - val_loss: 0.5972 - val_acc: 0.8924\n",
            "Epoch 73/100\n",
            "48000/48000 [==============================] - 17s 348us/step - loss: 0.0669 - acc: 0.9753 - val_loss: 0.7069 - val_acc: 0.8835\n",
            "Epoch 74/100\n",
            "48000/48000 [==============================] - 16s 343us/step - loss: 0.0712 - acc: 0.9751 - val_loss: 0.6214 - val_acc: 0.8962\n",
            "Epoch 75/100\n",
            "48000/48000 [==============================] - 17s 347us/step - loss: 0.0717 - acc: 0.9753 - val_loss: 0.5940 - val_acc: 0.8871\n",
            "Epoch 76/100\n",
            "48000/48000 [==============================] - 17s 346us/step - loss: 0.0611 - acc: 0.9786 - val_loss: 0.5822 - val_acc: 0.8982\n",
            "Epoch 77/100\n",
            "48000/48000 [==============================] - 17s 358us/step - loss: 0.0637 - acc: 0.9775 - val_loss: 0.6269 - val_acc: 0.8902\n",
            "Epoch 78/100\n",
            "48000/48000 [==============================] - 16s 337us/step - loss: 0.0664 - acc: 0.9772 - val_loss: 0.6508 - val_acc: 0.8923\n",
            "Epoch 79/100\n",
            "48000/48000 [==============================] - 16s 334us/step - loss: 0.0722 - acc: 0.9767 - val_loss: 0.6511 - val_acc: 0.8926\n",
            "Epoch 80/100\n",
            "48000/48000 [==============================] - 16s 339us/step - loss: 0.0627 - acc: 0.9774 - val_loss: 0.6239 - val_acc: 0.8932\n",
            "Epoch 81/100\n",
            "48000/48000 [==============================] - 16s 343us/step - loss: 0.0640 - acc: 0.9776 - val_loss: 0.6930 - val_acc: 0.8936\n",
            "Epoch 82/100\n",
            "48000/48000 [==============================] - 16s 339us/step - loss: 0.0610 - acc: 0.9778 - val_loss: 0.6188 - val_acc: 0.8936\n",
            "Epoch 83/100\n",
            "48000/48000 [==============================] - 16s 343us/step - loss: 0.0553 - acc: 0.9798 - val_loss: 0.6532 - val_acc: 0.8954\n",
            "Epoch 84/100\n",
            "48000/48000 [==============================] - 17s 355us/step - loss: 0.0605 - acc: 0.9784 - val_loss: 0.6791 - val_acc: 0.8903\n",
            "Epoch 85/100\n",
            "48000/48000 [==============================] - 18s 373us/step - loss: 0.0645 - acc: 0.9783 - val_loss: 0.6325 - val_acc: 0.8963\n",
            "Epoch 86/100\n",
            "48000/48000 [==============================] - 17s 364us/step - loss: 0.0544 - acc: 0.9810 - val_loss: 0.6512 - val_acc: 0.8978\n",
            "Epoch 87/100\n",
            "48000/48000 [==============================] - 18s 371us/step - loss: 0.0572 - acc: 0.9801 - val_loss: 0.6241 - val_acc: 0.8896\n",
            "Epoch 88/100\n",
            "48000/48000 [==============================] - 18s 381us/step - loss: 0.0561 - acc: 0.9801 - val_loss: 0.6623 - val_acc: 0.8923\n",
            "Epoch 89/100\n",
            "48000/48000 [==============================] - 18s 375us/step - loss: 0.0546 - acc: 0.9810 - val_loss: 0.6661 - val_acc: 0.8966\n",
            "Epoch 90/100\n",
            "48000/48000 [==============================] - 17s 357us/step - loss: 0.0534 - acc: 0.9811 - val_loss: 0.6805 - val_acc: 0.8928\n",
            "Epoch 91/100\n",
            "48000/48000 [==============================] - 17s 357us/step - loss: 0.0570 - acc: 0.9796 - val_loss: 0.6644 - val_acc: 0.8858\n",
            "Epoch 92/100\n",
            "48000/48000 [==============================] - 16s 343us/step - loss: 0.0597 - acc: 0.9799 - val_loss: 0.6457 - val_acc: 0.8955\n",
            "Epoch 93/100\n",
            "48000/48000 [==============================] - 16s 340us/step - loss: 0.0566 - acc: 0.9804 - val_loss: 0.8717 - val_acc: 0.8832\n",
            "Epoch 94/100\n",
            "48000/48000 [==============================] - 17s 344us/step - loss: 0.0624 - acc: 0.9794 - val_loss: 0.6562 - val_acc: 0.8935\n",
            "Epoch 95/100\n",
            "48000/48000 [==============================] - 18s 365us/step - loss: 0.0619 - acc: 0.9803 - val_loss: 0.5964 - val_acc: 0.8963\n",
            "Epoch 96/100\n",
            "48000/48000 [==============================] - 17s 355us/step - loss: 0.0500 - acc: 0.9825 - val_loss: 0.6086 - val_acc: 0.8944\n",
            "Epoch 97/100\n",
            "48000/48000 [==============================] - 18s 371us/step - loss: 0.0488 - acc: 0.9830 - val_loss: 0.7144 - val_acc: 0.8871\n",
            "Epoch 98/100\n",
            "48000/48000 [==============================] - 17s 355us/step - loss: 0.0576 - acc: 0.9816 - val_loss: 0.6770 - val_acc: 0.8921\n",
            "Epoch 99/100\n",
            "48000/48000 [==============================] - 17s 362us/step - loss: 0.0616 - acc: 0.9812 - val_loss: 0.5998 - val_acc: 0.8937\n",
            "Epoch 100/100\n",
            "48000/48000 [==============================] - 18s 374us/step - loss: 0.0664 - acc: 0.9795 - val_loss: 0.6081 - val_acc: 0.8988\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "92UxP1IGZGpf",
        "colab_type": "code",
        "outputId": "9446d147-e730-4a7e-b36d-11cccc23d309",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        }
      },
      "source": [
        "show_graph(history_1)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZIAAAEWCAYAAABMoxE0AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GpBsSot1PT2d",
        "colab_type": "code",
        "outputId": "92bb8bea-857e-464b-e661-cc203610e6c9",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "test_loss_1, test_acc_1 = model_1.evaluate(testX, testY)\n",
        "test_loss_1 , test_acc_1"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "10000/10000 [==============================] - 1s 73us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.671942839293927, 0.894)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "80Nsrt3cQRm1",
        "colab_type": "code",
        "outputId": "0c016eb7-4055-455c-f297-55acb4d0a5c3",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 255
        }
      },
      "source": [
        "model_2 = keras.Sequential([\n",
        "    keras.layers.Flatten(input_shape=(28,28)),\n",
        "    keras.layers.Dense(256, activation=tf.nn.tanh),\n",
        "    keras.layers.Dense(10, activation=tf.nn.softmax)\n",
        "])\n",
        "model_2.compile(optimizer='adam',\n",
        "              loss='sparse_categorical_crossentropy',\n",
        "              metrics=['accuracy'])\n",
        "model_2.summary()"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential_3\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "flatten_3 (Flatten)          (None, 784)               0         \n",
            "_________________________________________________________________\n",
            "dense_8 (Dense)              (None, 256)               200960    \n",
            "_________________________________________________________________\n",
            "dense_9 (Dense)              (None, 10)                2570      \n",
            "=================================================================\n",
            "Total params: 203,530\n",
            "Trainable params: 203,530\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ZNj1he1YR6tG",
        "colab_type": "code",
        "outputId": "6e01feed-4705-4b89-94b7-f6cee70d6683",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "history_2 = model_2.fit(trainX, trainY, epochs=epochs, validation_split=validation_split,verbose=verbose, shuffle = True)"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 48000 samples, validate on 12000 samples\n",
            "Epoch 1/100\n",
            "48000/48000 [==============================] - 8s 160us/step - loss: 0.4889 - acc: 0.8233 - val_loss: 0.3867 - val_acc: 0.8609\n",
            "Epoch 2/100\n",
            "48000/48000 [==============================] - 7s 155us/step - loss: 0.3821 - acc: 0.8604 - val_loss: 0.3800 - val_acc: 0.8625\n",
            "Epoch 3/100\n",
            "48000/48000 [==============================] - 8s 156us/step - loss: 0.3398 - acc: 0.8741 - val_loss: 0.3471 - val_acc: 0.8750\n",
            "Epoch 4/100\n",
            "48000/48000 [==============================] - 7s 156us/step - loss: 0.3167 - acc: 0.8826 - val_loss: 0.3430 - val_acc: 0.8723\n",
            "Epoch 5/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.2984 - acc: 0.8889 - val_loss: 0.3491 - val_acc: 0.8746\n",
            "Epoch 6/100\n",
            "48000/48000 [==============================] - 7s 154us/step - loss: 0.2831 - acc: 0.8939 - val_loss: 0.3356 - val_acc: 0.8792\n",
            "Epoch 7/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.2705 - acc: 0.8978 - val_loss: 0.3273 - val_acc: 0.8803\n",
            "Epoch 8/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.2571 - acc: 0.9038 - val_loss: 0.3210 - val_acc: 0.8842\n",
            "Epoch 9/100\n",
            "48000/48000 [==============================] - 7s 148us/step - loss: 0.2473 - acc: 0.9063 - val_loss: 0.3163 - val_acc: 0.8858\n",
            "Epoch 10/100\n",
            "48000/48000 [==============================] - 7s 156us/step - loss: 0.2382 - acc: 0.9101 - val_loss: 0.3327 - val_acc: 0.8808\n",
            "Epoch 11/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.2321 - acc: 0.9142 - val_loss: 0.3153 - val_acc: 0.8872\n",
            "Epoch 12/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.2236 - acc: 0.9157 - val_loss: 0.3105 - val_acc: 0.8915\n",
            "Epoch 13/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.2178 - acc: 0.9182 - val_loss: 0.2974 - val_acc: 0.8926\n",
            "Epoch 14/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.2069 - acc: 0.9225 - val_loss: 0.3327 - val_acc: 0.8813\n",
            "Epoch 15/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.2043 - acc: 0.9235 - val_loss: 0.2997 - val_acc: 0.8929\n",
            "Epoch 16/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.2002 - acc: 0.9260 - val_loss: 0.3148 - val_acc: 0.8924\n",
            "Epoch 17/100\n",
            "48000/48000 [==============================] - 7s 149us/step - loss: 0.1876 - acc: 0.9306 - val_loss: 0.3234 - val_acc: 0.8870\n",
            "Epoch 18/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.1852 - acc: 0.9311 - val_loss: 0.3132 - val_acc: 0.8911\n",
            "Epoch 19/100\n",
            "48000/48000 [==============================] - 8s 163us/step - loss: 0.1780 - acc: 0.9342 - val_loss: 0.3156 - val_acc: 0.8923\n",
            "Epoch 20/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.1775 - acc: 0.9338 - val_loss: 0.3136 - val_acc: 0.8902\n",
            "Epoch 21/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.1726 - acc: 0.9352 - val_loss: 0.3055 - val_acc: 0.8929\n",
            "Epoch 22/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.1671 - acc: 0.9388 - val_loss: 0.3257 - val_acc: 0.8902\n",
            "Epoch 23/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.1656 - acc: 0.9384 - val_loss: 0.3322 - val_acc: 0.8852\n",
            "Epoch 24/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.1637 - acc: 0.9394 - val_loss: 0.3277 - val_acc: 0.8886\n",
            "Epoch 25/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.1599 - acc: 0.9402 - val_loss: 0.3140 - val_acc: 0.8945\n",
            "Epoch 26/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.1548 - acc: 0.9416 - val_loss: 0.3393 - val_acc: 0.8894\n",
            "Epoch 27/100\n",
            "48000/48000 [==============================] - 7s 155us/step - loss: 0.1496 - acc: 0.9437 - val_loss: 0.3420 - val_acc: 0.8867\n",
            "Epoch 28/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.1466 - acc: 0.9463 - val_loss: 0.3299 - val_acc: 0.8938\n",
            "Epoch 29/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.1452 - acc: 0.9454 - val_loss: 0.3600 - val_acc: 0.8872\n",
            "Epoch 30/100\n",
            "48000/48000 [==============================] - 8s 166us/step - loss: 0.1430 - acc: 0.9459 - val_loss: 0.3404 - val_acc: 0.8892\n",
            "Epoch 31/100\n",
            "48000/48000 [==============================] - 7s 156us/step - loss: 0.1384 - acc: 0.9491 - val_loss: 0.3455 - val_acc: 0.8897\n",
            "Epoch 32/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.1382 - acc: 0.9482 - val_loss: 0.3603 - val_acc: 0.8897\n",
            "Epoch 33/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.1339 - acc: 0.9490 - val_loss: 0.3570 - val_acc: 0.8884\n",
            "Epoch 34/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.1317 - acc: 0.9517 - val_loss: 0.3707 - val_acc: 0.8878\n",
            "Epoch 35/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.1303 - acc: 0.9515 - val_loss: 0.3729 - val_acc: 0.8872\n",
            "Epoch 36/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.1240 - acc: 0.9543 - val_loss: 0.3536 - val_acc: 0.8906\n",
            "Epoch 37/100\n",
            "48000/48000 [==============================] - 7s 154us/step - loss: 0.1204 - acc: 0.9569 - val_loss: 0.3535 - val_acc: 0.8939\n",
            "Epoch 38/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.1192 - acc: 0.9546 - val_loss: 0.3639 - val_acc: 0.8908\n",
            "Epoch 39/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.1180 - acc: 0.9560 - val_loss: 0.3624 - val_acc: 0.8916\n",
            "Epoch 40/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.1162 - acc: 0.9577 - val_loss: 0.3659 - val_acc: 0.8892\n",
            "Epoch 41/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.1144 - acc: 0.9573 - val_loss: 0.3910 - val_acc: 0.8848\n",
            "Epoch 42/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.1117 - acc: 0.9593 - val_loss: 0.3872 - val_acc: 0.8868\n",
            "Epoch 43/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.1112 - acc: 0.9579 - val_loss: 0.3955 - val_acc: 0.8852\n",
            "Epoch 44/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.1093 - acc: 0.9588 - val_loss: 0.3837 - val_acc: 0.8870\n",
            "Epoch 45/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.1090 - acc: 0.9595 - val_loss: 0.3859 - val_acc: 0.8900\n",
            "Epoch 46/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.1023 - acc: 0.9623 - val_loss: 0.4021 - val_acc: 0.8848\n",
            "Epoch 47/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.1025 - acc: 0.9624 - val_loss: 0.4302 - val_acc: 0.8801\n",
            "Epoch 48/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.1023 - acc: 0.9633 - val_loss: 0.4010 - val_acc: 0.8887\n",
            "Epoch 49/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.0960 - acc: 0.9662 - val_loss: 0.4036 - val_acc: 0.8887\n",
            "Epoch 50/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.0953 - acc: 0.9649 - val_loss: 0.4040 - val_acc: 0.8875\n",
            "Epoch 51/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.0941 - acc: 0.9657 - val_loss: 0.4280 - val_acc: 0.8828\n",
            "Epoch 52/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.0945 - acc: 0.9655 - val_loss: 0.4352 - val_acc: 0.8814\n",
            "Epoch 53/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.0949 - acc: 0.9653 - val_loss: 0.4146 - val_acc: 0.8852\n",
            "Epoch 54/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.0913 - acc: 0.9669 - val_loss: 0.4131 - val_acc: 0.8869\n",
            "Epoch 55/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.0898 - acc: 0.9667 - val_loss: 0.4213 - val_acc: 0.8888\n",
            "Epoch 56/100\n",
            "48000/48000 [==============================] - 7s 156us/step - loss: 0.0873 - acc: 0.9678 - val_loss: 0.4299 - val_acc: 0.8872\n",
            "Epoch 57/100\n",
            "48000/48000 [==============================] - 7s 154us/step - loss: 0.0878 - acc: 0.9674 - val_loss: 0.4170 - val_acc: 0.8892\n",
            "Epoch 58/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.0869 - acc: 0.9674 - val_loss: 0.4508 - val_acc: 0.8833\n",
            "Epoch 59/100\n",
            "48000/48000 [==============================] - 7s 154us/step - loss: 0.0831 - acc: 0.9695 - val_loss: 0.4193 - val_acc: 0.8890\n",
            "Epoch 60/100\n",
            "48000/48000 [==============================] - 8s 159us/step - loss: 0.0796 - acc: 0.9709 - val_loss: 0.4576 - val_acc: 0.8800\n",
            "Epoch 61/100\n",
            "48000/48000 [==============================] - 8s 158us/step - loss: 0.0816 - acc: 0.9703 - val_loss: 0.4463 - val_acc: 0.8834\n",
            "Epoch 62/100\n",
            "48000/48000 [==============================] - 7s 154us/step - loss: 0.0797 - acc: 0.9704 - val_loss: 0.4310 - val_acc: 0.8885\n",
            "Epoch 63/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.0779 - acc: 0.9721 - val_loss: 0.4453 - val_acc: 0.8838\n",
            "Epoch 64/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.0807 - acc: 0.9701 - val_loss: 0.4643 - val_acc: 0.8825\n",
            "Epoch 65/100\n",
            "48000/48000 [==============================] - 7s 154us/step - loss: 0.0783 - acc: 0.9711 - val_loss: 0.4549 - val_acc: 0.8857\n",
            "Epoch 66/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.0745 - acc: 0.9725 - val_loss: 0.4685 - val_acc: 0.8792\n",
            "Epoch 67/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.0767 - acc: 0.9726 - val_loss: 0.4448 - val_acc: 0.8873\n",
            "Epoch 68/100\n",
            "48000/48000 [==============================] - 7s 155us/step - loss: 0.0719 - acc: 0.9741 - val_loss: 0.4612 - val_acc: 0.8862\n",
            "Epoch 69/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.0709 - acc: 0.9744 - val_loss: 0.5098 - val_acc: 0.8742\n",
            "Epoch 70/100\n",
            "48000/48000 [==============================] - 8s 163us/step - loss: 0.0716 - acc: 0.9735 - val_loss: 0.4651 - val_acc: 0.8854\n",
            "Epoch 71/100\n",
            "48000/48000 [==============================] - 8s 158us/step - loss: 0.0711 - acc: 0.9739 - val_loss: 0.4790 - val_acc: 0.8854\n",
            "Epoch 72/100\n",
            "48000/48000 [==============================] - 8s 167us/step - loss: 0.0669 - acc: 0.9755 - val_loss: 0.4690 - val_acc: 0.8853\n",
            "Epoch 73/100\n",
            "48000/48000 [==============================] - 8s 159us/step - loss: 0.0672 - acc: 0.9761 - val_loss: 0.4672 - val_acc: 0.8840\n",
            "Epoch 74/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.0664 - acc: 0.9761 - val_loss: 0.4820 - val_acc: 0.8822\n",
            "Epoch 75/100\n",
            "48000/48000 [==============================] - 7s 149us/step - loss: 0.0691 - acc: 0.9750 - val_loss: 0.4698 - val_acc: 0.8836\n",
            "Epoch 76/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.0679 - acc: 0.9751 - val_loss: 0.4715 - val_acc: 0.8873\n",
            "Epoch 77/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.0642 - acc: 0.9769 - val_loss: 0.4754 - val_acc: 0.8834\n",
            "Epoch 78/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.0636 - acc: 0.9760 - val_loss: 0.4847 - val_acc: 0.8843\n",
            "Epoch 79/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.0601 - acc: 0.9783 - val_loss: 0.5040 - val_acc: 0.8796\n",
            "Epoch 80/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.0646 - acc: 0.9765 - val_loss: 0.4868 - val_acc: 0.8850\n",
            "Epoch 81/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.0588 - acc: 0.9785 - val_loss: 0.4820 - val_acc: 0.8893\n",
            "Epoch 82/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.0613 - acc: 0.9780 - val_loss: 0.5003 - val_acc: 0.8827\n",
            "Epoch 83/100\n",
            "48000/48000 [==============================] - 7s 149us/step - loss: 0.0618 - acc: 0.9776 - val_loss: 0.5019 - val_acc: 0.8838\n",
            "Epoch 84/100\n",
            "48000/48000 [==============================] - 7s 148us/step - loss: 0.0621 - acc: 0.9777 - val_loss: 0.4965 - val_acc: 0.8812\n",
            "Epoch 85/100\n",
            "48000/48000 [==============================] - 7s 149us/step - loss: 0.0642 - acc: 0.9764 - val_loss: 0.5331 - val_acc: 0.8799\n",
            "Epoch 86/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.0645 - acc: 0.9764 - val_loss: 0.5030 - val_acc: 0.8844\n",
            "Epoch 87/100\n",
            "48000/48000 [==============================] - 7s 148us/step - loss: 0.0631 - acc: 0.9770 - val_loss: 0.4992 - val_acc: 0.8857\n",
            "Epoch 88/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.0627 - acc: 0.9771 - val_loss: 0.5332 - val_acc: 0.8796\n",
            "Epoch 89/100\n",
            "48000/48000 [==============================] - 7s 155us/step - loss: 0.0593 - acc: 0.9790 - val_loss: 0.4969 - val_acc: 0.8862\n",
            "Epoch 90/100\n",
            "48000/48000 [==============================] - 8s 162us/step - loss: 0.0592 - acc: 0.9783 - val_loss: 0.5187 - val_acc: 0.8840\n",
            "Epoch 91/100\n",
            "48000/48000 [==============================] - 8s 168us/step - loss: 0.0612 - acc: 0.9778 - val_loss: 0.5146 - val_acc: 0.8863\n",
            "Epoch 92/100\n",
            "48000/48000 [==============================] - 7s 154us/step - loss: 0.0573 - acc: 0.9786 - val_loss: 0.5176 - val_acc: 0.8848\n",
            "Epoch 93/100\n",
            "48000/48000 [==============================] - 7s 154us/step - loss: 0.0558 - acc: 0.9796 - val_loss: 0.5026 - val_acc: 0.8891\n",
            "Epoch 94/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.0561 - acc: 0.9798 - val_loss: 0.5210 - val_acc: 0.8838\n",
            "Epoch 95/100\n",
            "48000/48000 [==============================] - 7s 152us/step - loss: 0.0518 - acc: 0.9812 - val_loss: 0.5218 - val_acc: 0.8860\n",
            "Epoch 96/100\n",
            "48000/48000 [==============================] - 7s 150us/step - loss: 0.0505 - acc: 0.9819 - val_loss: 0.5479 - val_acc: 0.8838\n",
            "Epoch 97/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.0582 - acc: 0.9784 - val_loss: 0.5142 - val_acc: 0.8868\n",
            "Epoch 98/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.0517 - acc: 0.9811 - val_loss: 0.5275 - val_acc: 0.8863\n",
            "Epoch 99/100\n",
            "48000/48000 [==============================] - 7s 153us/step - loss: 0.0569 - acc: 0.9794 - val_loss: 0.5387 - val_acc: 0.8812\n",
            "Epoch 100/100\n",
            "48000/48000 [==============================] - 7s 151us/step - loss: 0.0542 - acc: 0.9799 - val_loss: 0.5223 - val_acc: 0.8875\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6S07aQc4aKwQ",
        "colab_type": "code",
        "outputId": "085b8e6d-ec53-4295-be37-5bbe98b0fd15",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 573
        }
      },
      "source": [
        "show_graph(history_2)"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n/5mVBpR5zvNjAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYsAAAEWCAYAAACXGLsWAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCxERiUnJQkREYlKyEBGRmP4fcyH4OB6h97AAAAAASUVORK5CYII=\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_Nxx_JpPSxyb",
        "colab_type": "code",
        "outputId": "9e512c84-8aa8-46e8-93c9-06691770ac74",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 51
        }
      },
      "source": [
        "test_loss_2, test_acc_2 = model_2.evaluate(testX, testY)\n",
        "test_loss_2, test_acc_2"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "10000/10000 [==============================] - 0s 38us/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(0.5894871108524501, 0.8785)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    }
  ]
}