neural-network/lenet-100/lenet100-mnist.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "LeNet100.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "UztzQmBgIIvH",
        "colab_type": "code",
        "outputId": "e93f2630-40ca-4250-a788-5a6ee29ec4ef",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 81
        }
      },
      "source": [
        "# libs import\n",
        "from keras.layers.convolutional import Convolution2D\n",
        "from keras.layers.convolutional import MaxPooling2D\n",
        "from keras.layers.core import Activation\n",
        "from keras.layers.core import Flatten\n",
        "from keras.layers.core import Dense\n",
        "from keras.models import Sequential\n",
        "from keras import models, layers\n",
        "from keras.optimizers import SGD\n",
        "from keras.utils import np_utils\n",
        "from sklearn.model_selection import train_test_split\n",
        "from sklearn.datasets import fetch_openml\n",
        "from tensorflow.python.util import deprecation\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "import keras"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Using TensorFlow backend.\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/html": [
              "<p style=\"color: red;\">\n",
              "The default version of TensorFlow in Colab will soon switch to TensorFlow 2.x.<br>\n",
              "We recommend you <a href=\"https://www.tensorflow.org/guide/migrate\" target=\"_blank\">upgrade</a> now \n",
              "or ensure your notebook will continue to use TensorFlow 1.x via the <code>%tensorflow_version 1.x</code> magic:\n",
              "<a href=\"https://colab.research.google.com/notebooks/tensorflow_version.ipynb\" target=\"_blank\">more info</a>.</p>\n"
            ],
            "text/plain": [
              "<IPython.core.display.HTML object>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-fGPINZtIP1D",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# fetching dataset\n",
        "dataset = fetch_openml('mnist_784')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OCJalnGnISTN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# reshaping\n",
        "data = dataset.data.reshape((dataset.data.shape[0], 28, 28))\n",
        "\n",
        "# reshaping + add new axis\n",
        "data = data[:, :, :, np.newaxis]"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LsNDtEUrIVXN",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# skalakan dari range 0-255 menjadi range 0-1\n",
        "scaled_data = data / 255.0"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EZ_5tuKlIXgQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# split training & test\n",
        "(x_train, x_test, y_train, y_test) = train_test_split(\n",
        "    scaled_data,\n",
        "    dataset.target.astype(\"float32\"), \n",
        "    test_size = 0.20)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IxsjuqfqIbam",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# transformasi target training dengan one-hot encoding\n",
        "y_train = np_utils.to_categorical(y_train, 10)\n",
        "\n",
        "# transformasi target testing dengan one-hot encoding\n",
        "y_test = np_utils.to_categorical(y_test, 10)"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "B8S8zMQXIdh6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# instantiasikan model sequential kosong\n",
        "model = Sequential()"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "E0TFTgMmKZsr",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# suppress warning\n",
        "deprecation._PRINT_DEPRECATION_WARNINGS = False"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "YzNcm2MxIe2E",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "import warnings\n",
        "warnings.filterwarnings('ignore')"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kPmaMohFIg0p",
        "colab_type": "code",
        "outputId": "0b3802a9-600d-42a0-d11f-f4129f1c84ac",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 238
        }
      },
      "source": [
        "# C1\n",
        "# Convolutional\n",
        "model.add(layers.Conv2D(6, kernel_size=(5, 5), strides=(1, 1), activation=\"tanh\", input_shape=(28,28,1), padding=\"same\"))\n",
        "\n",
        "# S2\n",
        "# Pooling\n",
        "model.add(layers.AveragePooling2D(pool_size=(2, 2), strides=(1, 1), padding=\"valid\"))\n",
        "\n",
        "# C3\n",
        "# Convolutional\n",
        "model.add(layers.Conv2D(16, kernel_size=(5, 5), strides=(1, 1), activation=\"tanh\", padding=\"valid\"))\n",
        "\n",
        "# S4\n",
        "# Pooling\n",
        "model.add(layers.AveragePooling2D(pool_size=(2, 2), strides=(2, 2), padding=\"valid\"))\n",
        "\n",
        "# C5\n",
        "# Fully Connected Convolutional\n",
        "model.add(layers.Conv2D(120, kernel_size=(5, 5), strides=(1, 1), activation=\"tanh\", padding=\"valid\"))\n",
        "\n",
        "#Flatten the CNN output so that we can connect it with fully connected layers\n",
        "model.add(layers.Flatten())\n",
        "\n",
        "# FC6 \n",
        "# Fully Connected\n",
        "model.add(layers.Dense(84, activation=\"tanh\"))\n",
        "\n",
        "# Output Layer (softmax activation)\n",
        "# Fully Connected\n",
        "model.add(layers.Dense(10, activation=\"softmax\"))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "error",
          "ename": "NameError",
          "evalue": "ignored",
          "traceback": [
            "\u001b[0;31m---------------------------------------------------------------------------\u001b[0m",
            "\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)",
            "\u001b[0;32m<ipython-input-1-3a83a868f3e1>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m()\u001b[0m\n\u001b[0;32m----> 1\u001b[0;31m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlayers\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mConv2D\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m6\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mkernel_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m5\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m5\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrides\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mactivation\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"tanh\"\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0minput_shape\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m28\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m28\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpadding\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"same\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      3\u001b[0m \u001b[0;31m# S2 Pooling Layer\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      4\u001b[0m \u001b[0mmodel\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0madd\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mlayers\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mAveragePooling2D\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mpool_size\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m2\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mstrides\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0;36m1\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpadding\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0;34m\"valid\"\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n",
            "\u001b[0;31mNameError\u001b[0m: name 'model' is not defined"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ERaHcxWVIiLZ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "# model compiling\n",
        "model.compile(\n",
        "    loss=keras.losses.categorical_crossentropy, \n",
        "    optimizer=\"SGD\", \n",
        "    metrics=[\"accuracy\"]\n",
        "    )"
      ],
      "execution_count": 0,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ctJf1bbAIkRh",
        "colab_type": "code",
        "outputId": "62129adc-526d-4576-dfc3-816d2b4ba2ff",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        }
      },
      "source": [
        "# epoch 100\n",
        "hist = model.fit(\n",
        "    x=x_train,y=y_train, \n",
        "    epochs=100, batch_size=128, \n",
        "    validation_data=(x_test, y_test), \n",
        "    verbose=1\n",
        "    )"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train on 56000 samples, validate on 14000 samples\n",
            "Epoch 1/100\n",
            "56000/56000 [==============================] - 3s 55us/step - loss: 0.7277 - acc: 0.8132 - val_loss: 0.3618 - val_acc: 0.8964\n",
            "Epoch 2/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.3119 - acc: 0.9121 - val_loss: 0.2789 - val_acc: 0.9195\n",
            "Epoch 3/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.2556 - acc: 0.9271 - val_loss: 0.2396 - val_acc: 0.9308\n",
            "Epoch 4/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.2229 - acc: 0.9368 - val_loss: 0.2120 - val_acc: 0.9385\n",
            "Epoch 5/100\n",
            "56000/56000 [==============================] - 3s 49us/step - loss: 0.1993 - acc: 0.9429 - val_loss: 0.1935 - val_acc: 0.9449\n",
            "Epoch 6/100\n",
            "56000/56000 [==============================] - 3s 49us/step - loss: 0.1814 - acc: 0.9480 - val_loss: 0.1801 - val_acc: 0.9489\n",
            "Epoch 7/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.1668 - acc: 0.9526 - val_loss: 0.1660 - val_acc: 0.9536\n",
            "Epoch 8/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.1545 - acc: 0.9562 - val_loss: 0.1535 - val_acc: 0.9568\n",
            "Epoch 9/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.1437 - acc: 0.9599 - val_loss: 0.1435 - val_acc: 0.9591\n",
            "Epoch 10/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.1342 - acc: 0.9624 - val_loss: 0.1347 - val_acc: 0.9619\n",
            "Epoch 11/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.1257 - acc: 0.9645 - val_loss: 0.1274 - val_acc: 0.9641\n",
            "Epoch 12/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.1184 - acc: 0.9665 - val_loss: 0.1215 - val_acc: 0.9645\n",
            "Epoch 13/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.1116 - acc: 0.9691 - val_loss: 0.1143 - val_acc: 0.9671\n",
            "Epoch 14/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.1051 - acc: 0.9711 - val_loss: 0.1097 - val_acc: 0.9678\n",
            "Epoch 15/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0997 - acc: 0.9726 - val_loss: 0.1041 - val_acc: 0.9701\n",
            "Epoch 16/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0948 - acc: 0.9737 - val_loss: 0.0987 - val_acc: 0.9705\n",
            "Epoch 17/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0900 - acc: 0.9754 - val_loss: 0.0943 - val_acc: 0.9728\n",
            "Epoch 18/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0857 - acc: 0.9766 - val_loss: 0.0906 - val_acc: 0.9739\n",
            "Epoch 19/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0816 - acc: 0.9779 - val_loss: 0.0872 - val_acc: 0.9741\n",
            "Epoch 20/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0781 - acc: 0.9787 - val_loss: 0.0838 - val_acc: 0.9752\n",
            "Epoch 21/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0747 - acc: 0.9797 - val_loss: 0.0806 - val_acc: 0.9759\n",
            "Epoch 22/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0714 - acc: 0.9806 - val_loss: 0.0785 - val_acc: 0.9770\n",
            "Epoch 23/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0687 - acc: 0.9814 - val_loss: 0.0757 - val_acc: 0.9779\n",
            "Epoch 24/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0658 - acc: 0.9824 - val_loss: 0.0735 - val_acc: 0.9776\n",
            "Epoch 25/100\n",
            "56000/56000 [==============================] - 3s 49us/step - loss: 0.0634 - acc: 0.9830 - val_loss: 0.0711 - val_acc: 0.9792\n",
            "Epoch 26/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0611 - acc: 0.9833 - val_loss: 0.0692 - val_acc: 0.9793\n",
            "Epoch 27/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0586 - acc: 0.9842 - val_loss: 0.0670 - val_acc: 0.9796\n",
            "Epoch 28/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0566 - acc: 0.9847 - val_loss: 0.0665 - val_acc: 0.9801\n",
            "Epoch 29/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0546 - acc: 0.9852 - val_loss: 0.0642 - val_acc: 0.9807\n",
            "Epoch 30/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0528 - acc: 0.9858 - val_loss: 0.0626 - val_acc: 0.9814\n",
            "Epoch 31/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0511 - acc: 0.9865 - val_loss: 0.0616 - val_acc: 0.9820\n",
            "Epoch 32/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0492 - acc: 0.9872 - val_loss: 0.0596 - val_acc: 0.9824\n",
            "Epoch 33/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0478 - acc: 0.9871 - val_loss: 0.0591 - val_acc: 0.9827\n",
            "Epoch 34/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0462 - acc: 0.9879 - val_loss: 0.0572 - val_acc: 0.9831\n",
            "Epoch 35/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0448 - acc: 0.9881 - val_loss: 0.0568 - val_acc: 0.9832\n",
            "Epoch 36/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0433 - acc: 0.9887 - val_loss: 0.0549 - val_acc: 0.9840\n",
            "Epoch 37/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0420 - acc: 0.9892 - val_loss: 0.0545 - val_acc: 0.9841\n",
            "Epoch 38/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0408 - acc: 0.9895 - val_loss: 0.0533 - val_acc: 0.9844\n",
            "Epoch 39/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0397 - acc: 0.9897 - val_loss: 0.0535 - val_acc: 0.9839\n",
            "Epoch 40/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0385 - acc: 0.9901 - val_loss: 0.0524 - val_acc: 0.9841\n",
            "Epoch 41/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0374 - acc: 0.9902 - val_loss: 0.0511 - val_acc: 0.9844\n",
            "Epoch 42/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0364 - acc: 0.9908 - val_loss: 0.0498 - val_acc: 0.9850\n",
            "Epoch 43/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0353 - acc: 0.9912 - val_loss: 0.0498 - val_acc: 0.9854\n",
            "Epoch 44/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0343 - acc: 0.9914 - val_loss: 0.0487 - val_acc: 0.9851\n",
            "Epoch 45/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0335 - acc: 0.9917 - val_loss: 0.0487 - val_acc: 0.9853\n",
            "Epoch 46/100\n",
            "56000/56000 [==============================] - 3s 45us/step - loss: 0.0326 - acc: 0.9918 - val_loss: 0.0485 - val_acc: 0.9851\n",
            "Epoch 47/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0318 - acc: 0.9919 - val_loss: 0.0464 - val_acc: 0.9856\n",
            "Epoch 48/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0308 - acc: 0.9926 - val_loss: 0.0461 - val_acc: 0.9864\n",
            "Epoch 49/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0302 - acc: 0.9926 - val_loss: 0.0453 - val_acc: 0.9863\n",
            "Epoch 50/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0294 - acc: 0.9931 - val_loss: 0.0452 - val_acc: 0.9858\n",
            "Epoch 51/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0287 - acc: 0.9932 - val_loss: 0.0446 - val_acc: 0.9863\n",
            "Epoch 52/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0279 - acc: 0.9935 - val_loss: 0.0470 - val_acc: 0.9856\n",
            "Epoch 53/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0273 - acc: 0.9936 - val_loss: 0.0434 - val_acc: 0.9868\n",
            "Epoch 54/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0265 - acc: 0.9939 - val_loss: 0.0435 - val_acc: 0.9870\n",
            "Epoch 55/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0259 - acc: 0.9941 - val_loss: 0.0431 - val_acc: 0.9865\n",
            "Epoch 56/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0253 - acc: 0.9941 - val_loss: 0.0432 - val_acc: 0.9865\n",
            "Epoch 57/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0247 - acc: 0.9944 - val_loss: 0.0421 - val_acc: 0.9871\n",
            "Epoch 58/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0241 - acc: 0.9945 - val_loss: 0.0416 - val_acc: 0.9875\n",
            "Epoch 59/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0236 - acc: 0.9948 - val_loss: 0.0413 - val_acc: 0.9874\n",
            "Epoch 60/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0230 - acc: 0.9950 - val_loss: 0.0413 - val_acc: 0.9876\n",
            "Epoch 61/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0226 - acc: 0.9950 - val_loss: 0.0410 - val_acc: 0.9871\n",
            "Epoch 62/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0219 - acc: 0.9952 - val_loss: 0.0407 - val_acc: 0.9875\n",
            "Epoch 63/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0215 - acc: 0.9953 - val_loss: 0.0401 - val_acc: 0.9876\n",
            "Epoch 64/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0210 - acc: 0.9956 - val_loss: 0.0398 - val_acc: 0.9876\n",
            "Epoch 65/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0206 - acc: 0.9958 - val_loss: 0.0402 - val_acc: 0.9871\n",
            "Epoch 66/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0201 - acc: 0.9958 - val_loss: 0.0396 - val_acc: 0.9872\n",
            "Epoch 67/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0197 - acc: 0.9960 - val_loss: 0.0395 - val_acc: 0.9879\n",
            "Epoch 68/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0193 - acc: 0.9961 - val_loss: 0.0389 - val_acc: 0.9885\n",
            "Epoch 69/100\n",
            "56000/56000 [==============================] - 3s 49us/step - loss: 0.0189 - acc: 0.9961 - val_loss: 0.0392 - val_acc: 0.9875\n",
            "Epoch 70/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0185 - acc: 0.9964 - val_loss: 0.0380 - val_acc: 0.9881\n",
            "Epoch 71/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0180 - acc: 0.9966 - val_loss: 0.0389 - val_acc: 0.9880\n",
            "Epoch 72/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0176 - acc: 0.9966 - val_loss: 0.0379 - val_acc: 0.9882\n",
            "Epoch 73/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0174 - acc: 0.9968 - val_loss: 0.0377 - val_acc: 0.9884\n",
            "Epoch 74/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0170 - acc: 0.9967 - val_loss: 0.0373 - val_acc: 0.9883\n",
            "Epoch 75/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0167 - acc: 0.9968 - val_loss: 0.0371 - val_acc: 0.9881\n",
            "Epoch 76/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0163 - acc: 0.9968 - val_loss: 0.0374 - val_acc: 0.9883\n",
            "Epoch 77/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0160 - acc: 0.9971 - val_loss: 0.0370 - val_acc: 0.9884\n",
            "Epoch 78/100\n",
            "56000/56000 [==============================] - 3s 45us/step - loss: 0.0157 - acc: 0.9971 - val_loss: 0.0370 - val_acc: 0.9881\n",
            "Epoch 79/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0154 - acc: 0.9974 - val_loss: 0.0362 - val_acc: 0.9891\n",
            "Epoch 80/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0150 - acc: 0.9976 - val_loss: 0.0365 - val_acc: 0.9884\n",
            "Epoch 81/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0148 - acc: 0.9976 - val_loss: 0.0362 - val_acc: 0.9885\n",
            "Epoch 82/100\n",
            "56000/56000 [==============================] - 3s 45us/step - loss: 0.0145 - acc: 0.9977 - val_loss: 0.0363 - val_acc: 0.9892\n",
            "Epoch 83/100\n",
            "56000/56000 [==============================] - 3s 45us/step - loss: 0.0142 - acc: 0.9976 - val_loss: 0.0362 - val_acc: 0.9886\n",
            "Epoch 84/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0139 - acc: 0.9977 - val_loss: 0.0356 - val_acc: 0.9889\n",
            "Epoch 85/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0137 - acc: 0.9978 - val_loss: 0.0354 - val_acc: 0.9885\n",
            "Epoch 86/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0133 - acc: 0.9978 - val_loss: 0.0357 - val_acc: 0.9889\n",
            "Epoch 87/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0131 - acc: 0.9981 - val_loss: 0.0355 - val_acc: 0.9886\n",
            "Epoch 88/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0129 - acc: 0.9981 - val_loss: 0.0352 - val_acc: 0.9891\n",
            "Epoch 89/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0127 - acc: 0.9981 - val_loss: 0.0351 - val_acc: 0.9889\n",
            "Epoch 90/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0124 - acc: 0.9981 - val_loss: 0.0350 - val_acc: 0.9892\n",
            "Epoch 91/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0122 - acc: 0.9983 - val_loss: 0.0350 - val_acc: 0.9891\n",
            "Epoch 92/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0120 - acc: 0.9983 - val_loss: 0.0344 - val_acc: 0.9892\n",
            "Epoch 93/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0117 - acc: 0.9984 - val_loss: 0.0348 - val_acc: 0.9891\n",
            "Epoch 94/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0116 - acc: 0.9984 - val_loss: 0.0354 - val_acc: 0.9891\n",
            "Epoch 95/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0114 - acc: 0.9984 - val_loss: 0.0343 - val_acc: 0.9893\n",
            "Epoch 96/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0112 - acc: 0.9983 - val_loss: 0.0343 - val_acc: 0.9891\n",
            "Epoch 97/100\n",
            "56000/56000 [==============================] - 3s 48us/step - loss: 0.0110 - acc: 0.9985 - val_loss: 0.0341 - val_acc: 0.9894\n",
            "Epoch 98/100\n",
            "56000/56000 [==============================] - 3s 46us/step - loss: 0.0108 - acc: 0.9986 - val_loss: 0.0345 - val_acc: 0.9893\n",
            "Epoch 99/100\n",
            "56000/56000 [==============================] - 3s 47us/step - loss: 0.0106 - acc: 0.9987 - val_loss: 0.0340 - val_acc: 0.9894\n",
            "Epoch 100/100\n",
            "56000/56000 [==============================] - 3s 50us/step - loss: 0.0104 - acc: 0.9985 - val_loss: 0.0341 - val_acc: 0.9889\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MRIb0Rc2IkkD",
        "colab_type": "code",
        "outputId": "5ab01dde-3273-470f-9657-7bed441a4a83",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 52
        }
      },
      "source": [
        "test_score = model.evaluate(x_test, y_test)\n",
        "print(\"Test loss {:.4f}, accuracy {:.2f}%\".format(test_score[0], test_score[1] * 100))"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "14000/14000 [==============================] - 1s 64us/step\n",
            "Test loss 0.0341, accuracy 98.89%\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kd4T7YnlIqO2",
        "colab_type": "code",
        "outputId": "fe7c6b5c-1684-4740-c32d-13b15c091a2c",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 312
        }
      },
      "source": [
        "#plotting\n",
        "f, ax = plt.subplots()\n",
        "ax.plot([None] + hist.history[\"acc\"])\n",
        "ax.plot([None] + hist.history[\"val_acc\"])\n",
        "\n",
        "ax.legend([\"Train acc\", \"Validation acc\"], loc = 0)\n",
        "ax.set_title(\"Training/Validation acc per Epoch\")\n",
        "ax.set_xlabel(\"Epoch\")\n",
        "ax.set_ylabel(\"acc\") "
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'acc')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAY4AAAEWCAYAAABxMXBSAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nXgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "nmTYvp6SIrzq",
        "colab_type": "code",
        "outputId": "3546dfad-f5c8-4ffe-cfe5-ad4bd2334aba",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 312
        }
      },
      "source": [
        "#plotting\n",
        "f, ax = plt.subplots()\n",
        "ax.plot([None] + hist.history[\"loss\"])\n",
        "ax.plot([None] + hist.history[\"val_loss\"])\n",
        "\n",
        "ax.legend([\"Train acc\", \"Validation acc\"], loc = 0)\n",
        "ax.set_title(\"Training/Validation acc per Epoch\")\n",
        "ax.set_xlabel(\"Epoch\")\n",
        "ax.set_ylabel(\"acc\")"
      ],
      "execution_count": 0,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'acc')"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    }
  ]
}