code/C4WK1A2_CNN_application_Keras.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "Goal: \n",
        "1. Create a mood classifier with TF Keras Sequential API\n",
        "2. Build a CNN to identify sign language with TF Keras Functional API"
      ],
      "metadata": {
        "id": "LVsBTDcqBgiV"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "NtBQPvYuBW6X",
        "outputId": "213ea1f7-a0a5-4113-c342-9a9498c6dc35"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Mounted at /gdrive\n",
            "/gdrive/My Drive/Coursera/Coursera_DeepLearningSpecialization/course4_coding_assignments/W1A2\n"
          ]
        }
      ],
      "source": [
        "# mount google drive. data and utils \n",
        "from google.colab import drive\n",
        "drive.mount('/gdrive')\n",
        "%cd /gdrive/My\\ Drive/Coursera/Coursera_DeepLearningSpecialization/course4_coding_assignments/W1A2/"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!ls"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "I-UhUWjGCAnT",
        "outputId": "22ea4b80-8713-4eeb-f90d-14aa02e9c4a1"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "C4WK1A2_CNN_application_Keras.ipynb    datasets\t\ttest_utils.py\n",
            "C4WK1A2_CNN_application_PyTorch.ipynb  images\t\tW1A2_CNN_structure.png\n",
            "cnn_utils.py\t\t\t       public_tests.py\n",
            "Convolution_model_Application.ipynb    __pycache__\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import math \n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt \n",
        "from matplotlib.pyplot import imread \n",
        "import scipy\n",
        "from PIL import Image \n",
        "import pandas as pd \n",
        "import tensorflow as tf\n",
        "import tensorflow.keras.layers as tfl\n",
        "from tensorflow.python.framework import ops\n",
        "from cnn_utils import * \n",
        "from test_utils import summary, comparator"
      ],
      "metadata": {
        "id": "UAgogj97CGqu"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "tf.__version__"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 35
        },
        "id": "ePXPdPXyC0Yi",
        "outputId": "f20d60f0-8df9-41b8-b43e-051230ca8ef8"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'2.11.0'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 4
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# load dataset\n",
        "X_train_orig, Y_train_orig, X_test_orig, Y_test_orig, classes = load_happy_dataset()\n",
        "\n",
        "# Normalize image vectors \n",
        "X_train = X_train_orig / 255.\n",
        "X_test = X_test_orig / 255. \n",
        "\n",
        "# Reshape \n",
        "Y_train = Y_train_orig.T\n",
        "Y_test = Y_test_orig.T \n",
        "\n",
        "print (\"number of training examples = \" + str(X_train.shape[0]))\n",
        "print (\"number of test examples = \" + str(X_test.shape[0]))\n",
        "print (\"X_train shape: \" + str(X_train.shape))\n",
        "print (\"Y_train shape: \" + str(Y_train.shape))\n",
        "print (\"X_test shape: \" + str(X_test.shape))\n",
        "print (\"Y_test shape: \" + str(Y_test.shape))\n",
        "print(f\"number of classes: {len(classes)} {classes}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "x6sOaCAzC0uF",
        "outputId": "df4380fe-91b3-4968-ab73-bda8007be07c"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "number of training examples = 600\n",
            "number of test examples = 150\n",
            "X_train shape: (600, 64, 64, 3)\n",
            "Y_train shape: (600, 1)\n",
            "X_test shape: (150, 64, 64, 3)\n",
            "Y_test shape: (150, 1)\n",
            "number of classes: 2 [0 1]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "index = 12\n",
        "plt.imshow(X_train_orig[index])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 268
        },
        "id": "A09g83emEu9m",
        "outputId": "074691c5-5e69-459c-a803-f16d6c46bb4e"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Sequential Model\n",
        "\n",
        "Sequential API is ideal for building models where each layer has exactly one input tensor and one output tensor. It's only appropriate for simple, straightforward tasks. "
      ],
      "metadata": {
        "id": "etKs1wVHJUTW"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def happyModel():\n",
        "  \"\"\"\n",
        "  Returns: \n",
        "  TF Keras model \n",
        "  \"\"\"\n",
        "\n",
        "  model = tf.keras.Sequential([\n",
        "                               tf.keras.layers.ZeroPadding2D(padding=(3,3), input_shape=(64,64,3), data_format='channels_last'),\n",
        "                               tf.keras.layers.Conv2D(filters=32, kernel_size=(7,7), strides=1),\n",
        "                               tf.keras.layers.BatchNormalization(axis=3),\n",
        "                               tf.keras.layers.ReLU(),\n",
        "                               tf.keras.layers.MaxPool2D(),\n",
        "                               tf.keras.layers.Flatten(),\n",
        "                               tf.keras.layers.Dense(units=1, activation='sigmoid')\n",
        "  ])\n",
        "  return model"
      ],
      "metadata": {
        "id": "nvdLI0o7E7cm"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "happy_model = happyModel()\n",
        "for layer in summary(happy_model):\n",
        "  print(layer)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "omgc6ZkZM5DW",
        "outputId": "a16bb24a-d2cd-4927-9cd1-35d1ed822efb"
      },
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "['ZeroPadding2D', (None, 70, 70, 3), 0, ((3, 3), (3, 3))]\n",
            "['Conv2D', (None, 64, 64, 32), 4736, 'valid', 'linear', 'GlorotUniform']\n",
            "['BatchNormalization', (None, 64, 64, 32), 128]\n",
            "['ReLU', (None, 64, 64, 32), 0]\n",
            "['MaxPooling2D', (None, 32, 32, 32), 0, (2, 2), (2, 2), 'valid']\n",
            "['Flatten', (None, 32768), 0]\n",
            "['Dense', (None, 1), 32769, 'sigmoid']\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "happy_model.compile(optimizer='adam', loss='binary_crossentropy',metrics=['accuracy'])"
      ],
      "metadata": {
        "id": "dzhCJZn8NLzE"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "happy_model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "atbua8xyNfAg",
        "outputId": "66ba7a50-b301-473c-e3b9-b675793e62f5"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " zero_padding2d (ZeroPadding  (None, 70, 70, 3)        0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d (Conv2D)             (None, 64, 64, 32)        4736      \n",
            "                                                                 \n",
            " batch_normalization (BatchN  (None, 64, 64, 32)       128       \n",
            " ormalization)                                                   \n",
            "                                                                 \n",
            " re_lu (ReLU)                (None, 64, 64, 32)        0         \n",
            "                                                                 \n",
            " max_pooling2d (MaxPooling2D  (None, 32, 32, 32)       0         \n",
            " )                                                               \n",
            "                                                                 \n",
            " flatten (Flatten)           (None, 32768)             0         \n",
            "                                                                 \n",
            " dense (Dense)               (None, 1)                 32769     \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 37,633\n",
            "Trainable params: 37,569\n",
            "Non-trainable params: 64\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "happy_model.fit(X_train, Y_train, epochs = 10, batch_size = 16)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TL2mmxXtNhmP",
        "outputId": "af20305e-aef8-4009-be20-7ca8796939f9"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "38/38 [==============================] - 7s 127ms/step - loss: 0.9047 - accuracy: 0.7567\n",
            "Epoch 2/10\n",
            "38/38 [==============================] - 5s 129ms/step - loss: 0.2251 - accuracy: 0.9133\n",
            "Epoch 3/10\n",
            "38/38 [==============================] - 7s 176ms/step - loss: 0.1291 - accuracy: 0.9483\n",
            "Epoch 4/10\n",
            "38/38 [==============================] - 5s 122ms/step - loss: 0.1057 - accuracy: 0.9567\n",
            "Epoch 5/10\n",
            "38/38 [==============================] - 6s 165ms/step - loss: 0.1917 - accuracy: 0.9333\n",
            "Epoch 6/10\n",
            "38/38 [==============================] - 5s 142ms/step - loss: 0.0790 - accuracy: 0.9767\n",
            "Epoch 7/10\n",
            "38/38 [==============================] - 5s 122ms/step - loss: 0.1178 - accuracy: 0.9550\n",
            "Epoch 8/10\n",
            "38/38 [==============================] - 7s 183ms/step - loss: 0.2065 - accuracy: 0.9400\n",
            "Epoch 9/10\n",
            "38/38 [==============================] - 5s 122ms/step - loss: 0.0707 - accuracy: 0.9767\n",
            "Epoch 10/10\n",
            "38/38 [==============================] - 5s 124ms/step - loss: 0.1561 - accuracy: 0.9400\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.callbacks.History at 0x7f79edc3e2e0>"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "happy_model.evaluate(X_test, Y_test)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "cBS6lmEhNptK",
        "outputId": "bb70f9c7-10b8-47cd-f74c-388db3a5bbca"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "5/5 [==============================] - 1s 65ms/step - loss: 0.5702 - accuracy: 0.8000\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.5701653957366943, 0.800000011920929]"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Functional API\n",
        "\n",
        "Functional API can handle models with non-linear topology, shared layers, and layers with multiple inputs or outputs. \n"
      ],
      "metadata": {
        "id": "Lp1jwSvhPZ81"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "X_train_orig, Y_train_orig, X_test_orig, Y_test_orig, classes = load_signs_dataset()"
      ],
      "metadata": {
        "id": "DpBmA630Oclo"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "index = 13\n",
        "plt.imshow(X_train_orig[index])\n",
        "print(f\"y = {np.squeeze(Y_train_orig[:, index])}\")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 285
        },
        "id": "Uxvm6ayyRX39",
        "outputId": "2e65a936-f273-46b2-bd0d-3f92692670eb"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "y = 5\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "X_train = X_train_orig/255.\n",
        "X_test = X_test_orig/255.\n",
        "Y_train = convert_to_one_hot(Y_train_orig, 6).T\n",
        "Y_test = convert_to_one_hot(Y_test_orig, 6).T\n",
        "print (\"number of training examples = \" + str(X_train.shape[0]))\n",
        "print (\"number of test examples = \" + str(X_test.shape[0]))\n",
        "print (\"X_train shape: \" + str(X_train.shape))\n",
        "print (\"Y_train shape: \" + str(Y_train.shape))\n",
        "print (\"X_test shape: \" + str(X_test.shape))\n",
        "print (\"Y_test shape: \" + str(Y_test.shape))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nzbB7An0T9IT",
        "outputId": "85eefec6-dabb-443a-d910-e38bc18c38c3"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "number of training examples = 1080\n",
            "number of test examples = 120\n",
            "X_train shape: (1080, 64, 64, 3)\n",
            "Y_train shape: (1080, 6)\n",
            "X_test shape: (120, 64, 64, 3)\n",
            "Y_test shape: (120, 6)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Conv_model\n",
        "- Conv2D: 4 * 4 filter * 8 + padding = \"same\"\n",
        "- ReLU\n",
        "- MaxPool2D: 8 * 8 filter + 8 * 8 stride + padding = \"same\"\n",
        "- Conv2D: 2 * 2 filter * 16 + stride = 1 + padding = \"same\"\n",
        "- ReLU\n",
        "- MaxPool2D: 4 * 4 filter + 4 * 4 stride + padding = \"same\"\n",
        "- Flatten\n",
        "- Full-connected (Dense): 6 neurons + softmax"
      ],
      "metadata": {
        "id": "bnkvrkDZVWDw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def convolutional_model(input_shape):\n",
        "  \"\"\"\n",
        "  Returns: model\n",
        "  \"\"\"\n",
        "  input_img = tf.keras.Input(shape=input_shape)\n",
        "  Z1 = tf.keras.layers.Conv2D(filters=8, kernel_size=(4,4), padding=\"same\")(input_img)\n",
        "  A1 = tf.keras.layers.ReLU()(Z1)\n",
        "  P1 = tf.keras.layers.MaxPool2D(pool_size=(8,8), strides=(8,8), padding=\"same\")(A1)\n",
        "  Z2 = tf.keras.layers.Conv2D(filters=16, kernel_size=(2,2), padding=\"same\")(P1)\n",
        "  A2 = tf.keras.layers.ReLU()(Z2)\n",
        "  P2 = tf.keras.layers.MaxPool2D(pool_size=(4,4), strides=(4,4), padding=\"same\")(A2)\n",
        "  F = tf.keras.layers.Flatten()(P2)\n",
        "  outputs = tf.keras.layers.Dense(units=6, activation='softmax')(F)\n",
        "\n",
        "  model = tf.keras.Model(inputs=input_img, outputs = outputs)\n",
        "  return model"
      ],
      "metadata": {
        "id": "bpJZ-HG9UCGz"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "conv_model = convolutional_model((64,64,3))\n",
        "conv_model.compile(optimizer='adam',\n",
        "                  loss='categorical_crossentropy',\n",
        "                  metrics=['accuracy'])\n",
        "conv_model.summary()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "m4x2LyAtWDaB",
        "outputId": "aaff6533-0a3a-4da9-bc00-6a25ffcde957"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_2 (InputLayer)        [(None, 64, 64, 3)]       0         \n",
            "                                                                 \n",
            " conv2d_3 (Conv2D)           (None, 64, 64, 8)         392       \n",
            "                                                                 \n",
            " re_lu_3 (ReLU)              (None, 64, 64, 8)         0         \n",
            "                                                                 \n",
            " max_pooling2d_3 (MaxPooling  (None, 8, 8, 8)          0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_4 (Conv2D)           (None, 8, 8, 16)          528       \n",
            "                                                                 \n",
            " re_lu_4 (ReLU)              (None, 8, 8, 16)          0         \n",
            "                                                                 \n",
            " max_pooling2d_4 (MaxPooling  (None, 2, 2, 16)         0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " flatten_2 (Flatten)         (None, 64)                0         \n",
            "                                                                 \n",
            " dense_2 (Dense)             (None, 6)                 390       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 1,310\n",
            "Trainable params: 1,310\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "train_dataset = tf.data.Dataset.from_tensor_slices((X_train, Y_train)).batch(64)\n",
        "test_dataset = tf.data.Dataset.from_tensor_slices((X_test, Y_test)).batch(64)\n",
        "history = conv_model.fit(train_dataset, epochs = 100, validation_data = test_dataset)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "85XaQj2iXCk3",
        "outputId": "212ffc3d-47dc-4df3-8d5d-cd0e173e93fa"
      },
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "17/17 [==============================] - 3s 136ms/step - loss: 1.7971 - accuracy: 0.1778 - val_loss: 1.7872 - val_accuracy: 0.1750\n",
            "Epoch 2/100\n",
            "17/17 [==============================] - 2s 133ms/step - loss: 1.7752 - accuracy: 0.2269 - val_loss: 1.7804 - val_accuracy: 0.1667\n",
            "Epoch 3/100\n",
            "17/17 [==============================] - 2s 148ms/step - loss: 1.7668 - accuracy: 0.2694 - val_loss: 1.7737 - val_accuracy: 0.2667\n",
            "Epoch 4/100\n",
            "17/17 [==============================] - 4s 228ms/step - loss: 1.7570 - accuracy: 0.3157 - val_loss: 1.7673 - val_accuracy: 0.2750\n",
            "Epoch 5/100\n",
            "17/17 [==============================] - 2s 128ms/step - loss: 1.7480 - accuracy: 0.3194 - val_loss: 1.7607 - val_accuracy: 0.2833\n",
            "Epoch 6/100\n",
            "17/17 [==============================] - 2s 123ms/step - loss: 1.7352 - accuracy: 0.3380 - val_loss: 1.7483 - val_accuracy: 0.3000\n",
            "Epoch 7/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 1.7142 - accuracy: 0.3704 - val_loss: 1.7308 - val_accuracy: 0.2833\n",
            "Epoch 8/100\n",
            "17/17 [==============================] - 2s 130ms/step - loss: 1.6877 - accuracy: 0.3889 - val_loss: 1.7074 - val_accuracy: 0.3083\n",
            "Epoch 9/100\n",
            "17/17 [==============================] - 3s 177ms/step - loss: 1.6539 - accuracy: 0.3954 - val_loss: 1.6790 - val_accuracy: 0.3167\n",
            "Epoch 10/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 1.6123 - accuracy: 0.4157 - val_loss: 1.6475 - val_accuracy: 0.3083\n",
            "Epoch 11/100\n",
            "17/17 [==============================] - 2s 128ms/step - loss: 1.5683 - accuracy: 0.4324 - val_loss: 1.6132 - val_accuracy: 0.3167\n",
            "Epoch 12/100\n",
            "17/17 [==============================] - 2s 127ms/step - loss: 1.5212 - accuracy: 0.4546 - val_loss: 1.5750 - val_accuracy: 0.3333\n",
            "Epoch 13/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 1.4726 - accuracy: 0.4833 - val_loss: 1.5378 - val_accuracy: 0.3333\n",
            "Epoch 14/100\n",
            "17/17 [==============================] - 4s 221ms/step - loss: 1.4223 - accuracy: 0.5139 - val_loss: 1.4893 - val_accuracy: 0.3917\n",
            "Epoch 15/100\n",
            "17/17 [==============================] - 3s 205ms/step - loss: 1.3725 - accuracy: 0.5417 - val_loss: 1.4424 - val_accuracy: 0.4250\n",
            "Epoch 16/100\n",
            "17/17 [==============================] - 2s 127ms/step - loss: 1.3263 - accuracy: 0.5611 - val_loss: 1.3965 - val_accuracy: 0.4500\n",
            "Epoch 17/100\n",
            "17/17 [==============================] - 2s 128ms/step - loss: 1.2831 - accuracy: 0.5685 - val_loss: 1.3536 - val_accuracy: 0.4750\n",
            "Epoch 18/100\n",
            "17/17 [==============================] - 4s 229ms/step - loss: 1.2410 - accuracy: 0.5769 - val_loss: 1.3091 - val_accuracy: 0.5000\n",
            "Epoch 19/100\n",
            "17/17 [==============================] - 2s 126ms/step - loss: 1.2011 - accuracy: 0.5889 - val_loss: 1.2644 - val_accuracy: 0.5167\n",
            "Epoch 20/100\n",
            "17/17 [==============================] - 2s 125ms/step - loss: 1.1631 - accuracy: 0.5991 - val_loss: 1.2247 - val_accuracy: 0.5250\n",
            "Epoch 21/100\n",
            "17/17 [==============================] - 2s 128ms/step - loss: 1.1256 - accuracy: 0.6102 - val_loss: 1.1835 - val_accuracy: 0.5417\n",
            "Epoch 22/100\n",
            "17/17 [==============================] - 2s 138ms/step - loss: 1.0913 - accuracy: 0.6231 - val_loss: 1.1474 - val_accuracy: 0.5583\n",
            "Epoch 23/100\n",
            "17/17 [==============================] - 4s 221ms/step - loss: 1.0586 - accuracy: 0.6315 - val_loss: 1.1136 - val_accuracy: 0.5667\n",
            "Epoch 24/100\n",
            "17/17 [==============================] - 3s 145ms/step - loss: 1.0270 - accuracy: 0.6417 - val_loss: 1.0791 - val_accuracy: 0.5667\n",
            "Epoch 25/100\n",
            "17/17 [==============================] - 2s 128ms/step - loss: 0.9981 - accuracy: 0.6537 - val_loss: 1.0489 - val_accuracy: 0.5750\n",
            "Epoch 26/100\n",
            "17/17 [==============================] - 2s 126ms/step - loss: 0.9702 - accuracy: 0.6713 - val_loss: 1.0198 - val_accuracy: 0.5833\n",
            "Epoch 27/100\n",
            "17/17 [==============================] - 2s 128ms/step - loss: 0.9432 - accuracy: 0.6824 - val_loss: 0.9909 - val_accuracy: 0.6250\n",
            "Epoch 28/100\n",
            "17/17 [==============================] - 3s 170ms/step - loss: 0.9179 - accuracy: 0.6944 - val_loss: 0.9633 - val_accuracy: 0.6250\n",
            "Epoch 29/100\n",
            "17/17 [==============================] - 4s 209ms/step - loss: 0.8942 - accuracy: 0.7000 - val_loss: 0.9363 - val_accuracy: 0.6417\n",
            "Epoch 30/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.8705 - accuracy: 0.7120 - val_loss: 0.9097 - val_accuracy: 0.6667\n",
            "Epoch 31/100\n",
            "17/17 [==============================] - 2s 127ms/step - loss: 0.8475 - accuracy: 0.7185 - val_loss: 0.8840 - val_accuracy: 0.6583\n",
            "Epoch 32/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 0.8254 - accuracy: 0.7241 - val_loss: 0.8600 - val_accuracy: 0.6667\n",
            "Epoch 33/100\n",
            "17/17 [==============================] - 3s 200ms/step - loss: 0.8037 - accuracy: 0.7389 - val_loss: 0.8369 - val_accuracy: 0.6917\n",
            "Epoch 34/100\n",
            "17/17 [==============================] - 2s 122ms/step - loss: 0.7834 - accuracy: 0.7435 - val_loss: 0.8161 - val_accuracy: 0.7083\n",
            "Epoch 35/100\n",
            "17/17 [==============================] - 2s 123ms/step - loss: 0.7641 - accuracy: 0.7500 - val_loss: 0.7958 - val_accuracy: 0.7417\n",
            "Epoch 36/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.7453 - accuracy: 0.7565 - val_loss: 0.7768 - val_accuracy: 0.7500\n",
            "Epoch 37/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 0.7274 - accuracy: 0.7657 - val_loss: 0.7589 - val_accuracy: 0.7583\n",
            "Epoch 38/100\n",
            "17/17 [==============================] - 3s 207ms/step - loss: 0.7100 - accuracy: 0.7741 - val_loss: 0.7414 - val_accuracy: 0.7583\n",
            "Epoch 39/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 0.6934 - accuracy: 0.7843 - val_loss: 0.7245 - val_accuracy: 0.7750\n",
            "Epoch 40/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 0.6776 - accuracy: 0.7889 - val_loss: 0.7086 - val_accuracy: 0.7667\n",
            "Epoch 41/100\n",
            "17/17 [==============================] - 2s 130ms/step - loss: 0.6628 - accuracy: 0.7917 - val_loss: 0.6933 - val_accuracy: 0.7667\n",
            "Epoch 42/100\n",
            "17/17 [==============================] - 2s 133ms/step - loss: 0.6482 - accuracy: 0.7972 - val_loss: 0.6787 - val_accuracy: 0.7750\n",
            "Epoch 43/100\n",
            "17/17 [==============================] - 4s 218ms/step - loss: 0.6347 - accuracy: 0.8019 - val_loss: 0.6650 - val_accuracy: 0.7750\n",
            "Epoch 44/100\n",
            "17/17 [==============================] - 2s 135ms/step - loss: 0.6218 - accuracy: 0.8102 - val_loss: 0.6519 - val_accuracy: 0.7750\n",
            "Epoch 45/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 0.6094 - accuracy: 0.8157 - val_loss: 0.6387 - val_accuracy: 0.7917\n",
            "Epoch 46/100\n",
            "17/17 [==============================] - 2s 126ms/step - loss: 0.5971 - accuracy: 0.8194 - val_loss: 0.6271 - val_accuracy: 0.7833\n",
            "Epoch 47/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 0.5852 - accuracy: 0.8278 - val_loss: 0.6150 - val_accuracy: 0.7917\n",
            "Epoch 48/100\n",
            "17/17 [==============================] - 3s 176ms/step - loss: 0.5738 - accuracy: 0.8324 - val_loss: 0.6045 - val_accuracy: 0.8000\n",
            "Epoch 49/100\n",
            "17/17 [==============================] - 3s 202ms/step - loss: 0.5630 - accuracy: 0.8361 - val_loss: 0.5935 - val_accuracy: 0.8000\n",
            "Epoch 50/100\n",
            "17/17 [==============================] - 2s 125ms/step - loss: 0.5528 - accuracy: 0.8407 - val_loss: 0.5835 - val_accuracy: 0.8000\n",
            "Epoch 51/100\n",
            "17/17 [==============================] - 2s 125ms/step - loss: 0.5431 - accuracy: 0.8426 - val_loss: 0.5738 - val_accuracy: 0.8000\n",
            "Epoch 52/100\n",
            "17/17 [==============================] - 2s 127ms/step - loss: 0.5335 - accuracy: 0.8463 - val_loss: 0.5640 - val_accuracy: 0.8083\n",
            "Epoch 53/100\n",
            "17/17 [==============================] - 3s 181ms/step - loss: 0.5244 - accuracy: 0.8491 - val_loss: 0.5545 - val_accuracy: 0.8083\n",
            "Epoch 54/100\n",
            "17/17 [==============================] - 2s 126ms/step - loss: 0.5156 - accuracy: 0.8500 - val_loss: 0.5454 - val_accuracy: 0.8083\n",
            "Epoch 55/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 0.5073 - accuracy: 0.8546 - val_loss: 0.5369 - val_accuracy: 0.8083\n",
            "Epoch 56/100\n",
            "17/17 [==============================] - 2s 126ms/step - loss: 0.4994 - accuracy: 0.8537 - val_loss: 0.5288 - val_accuracy: 0.8083\n",
            "Epoch 57/100\n",
            "17/17 [==============================] - 2s 125ms/step - loss: 0.4919 - accuracy: 0.8556 - val_loss: 0.5205 - val_accuracy: 0.8167\n",
            "Epoch 58/100\n",
            "17/17 [==============================] - 3s 199ms/step - loss: 0.4845 - accuracy: 0.8574 - val_loss: 0.5132 - val_accuracy: 0.8250\n",
            "Epoch 59/100\n",
            "17/17 [==============================] - 3s 181ms/step - loss: 0.4771 - accuracy: 0.8602 - val_loss: 0.5059 - val_accuracy: 0.8333\n",
            "Epoch 60/100\n",
            "17/17 [==============================] - 2s 130ms/step - loss: 0.4705 - accuracy: 0.8611 - val_loss: 0.4990 - val_accuracy: 0.8333\n",
            "Epoch 61/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 0.4639 - accuracy: 0.8648 - val_loss: 0.4927 - val_accuracy: 0.8333\n",
            "Epoch 62/100\n",
            "17/17 [==============================] - 2s 127ms/step - loss: 0.4575 - accuracy: 0.8657 - val_loss: 0.4872 - val_accuracy: 0.8333\n",
            "Epoch 63/100\n",
            "17/17 [==============================] - 4s 225ms/step - loss: 0.4518 - accuracy: 0.8685 - val_loss: 0.4809 - val_accuracy: 0.8333\n",
            "Epoch 64/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.4461 - accuracy: 0.8676 - val_loss: 0.4761 - val_accuracy: 0.8333\n",
            "Epoch 65/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 0.4401 - accuracy: 0.8704 - val_loss: 0.4696 - val_accuracy: 0.8333\n",
            "Epoch 66/100\n",
            "17/17 [==============================] - 2s 125ms/step - loss: 0.4347 - accuracy: 0.8704 - val_loss: 0.4648 - val_accuracy: 0.8333\n",
            "Epoch 67/100\n",
            "17/17 [==============================] - 2s 128ms/step - loss: 0.4295 - accuracy: 0.8750 - val_loss: 0.4594 - val_accuracy: 0.8333\n",
            "Epoch 68/100\n",
            "17/17 [==============================] - 4s 220ms/step - loss: 0.4243 - accuracy: 0.8778 - val_loss: 0.4550 - val_accuracy: 0.8333\n",
            "Epoch 69/100\n",
            "17/17 [==============================] - 2s 133ms/step - loss: 0.4197 - accuracy: 0.8796 - val_loss: 0.4497 - val_accuracy: 0.8417\n",
            "Epoch 70/100\n",
            "17/17 [==============================] - 2s 128ms/step - loss: 0.4145 - accuracy: 0.8787 - val_loss: 0.4453 - val_accuracy: 0.8417\n",
            "Epoch 71/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.4093 - accuracy: 0.8806 - val_loss: 0.4407 - val_accuracy: 0.8500\n",
            "Epoch 72/100\n",
            "17/17 [==============================] - 2s 130ms/step - loss: 0.4048 - accuracy: 0.8806 - val_loss: 0.4359 - val_accuracy: 0.8500\n",
            "Epoch 73/100\n",
            "17/17 [==============================] - 3s 198ms/step - loss: 0.4002 - accuracy: 0.8824 - val_loss: 0.4315 - val_accuracy: 0.8500\n",
            "Epoch 74/100\n",
            "17/17 [==============================] - 4s 209ms/step - loss: 0.3955 - accuracy: 0.8852 - val_loss: 0.4276 - val_accuracy: 0.8500\n",
            "Epoch 75/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.3911 - accuracy: 0.8833 - val_loss: 0.4238 - val_accuracy: 0.8500\n",
            "Epoch 76/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.3870 - accuracy: 0.8833 - val_loss: 0.4200 - val_accuracy: 0.8500\n",
            "Epoch 77/100\n",
            "17/17 [==============================] - 4s 225ms/step - loss: 0.3832 - accuracy: 0.8843 - val_loss: 0.4166 - val_accuracy: 0.8500\n",
            "Epoch 78/100\n",
            "17/17 [==============================] - 2s 125ms/step - loss: 0.3789 - accuracy: 0.8870 - val_loss: 0.4135 - val_accuracy: 0.8500\n",
            "Epoch 79/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.3751 - accuracy: 0.8861 - val_loss: 0.4099 - val_accuracy: 0.8500\n",
            "Epoch 80/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.3714 - accuracy: 0.8898 - val_loss: 0.4070 - val_accuracy: 0.8500\n",
            "Epoch 81/100\n",
            "17/17 [==============================] - 3s 167ms/step - loss: 0.3677 - accuracy: 0.8907 - val_loss: 0.4034 - val_accuracy: 0.8500\n",
            "Epoch 82/100\n",
            "17/17 [==============================] - 2s 126ms/step - loss: 0.3640 - accuracy: 0.8926 - val_loss: 0.4003 - val_accuracy: 0.8500\n",
            "Epoch 83/100\n",
            "17/17 [==============================] - 2s 128ms/step - loss: 0.3606 - accuracy: 0.8926 - val_loss: 0.3970 - val_accuracy: 0.8500\n",
            "Epoch 84/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 0.3574 - accuracy: 0.8954 - val_loss: 0.3938 - val_accuracy: 0.8500\n",
            "Epoch 85/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.3543 - accuracy: 0.8954 - val_loss: 0.3905 - val_accuracy: 0.8500\n",
            "Epoch 86/100\n",
            "17/17 [==============================] - 4s 225ms/step - loss: 0.3510 - accuracy: 0.8954 - val_loss: 0.3882 - val_accuracy: 0.8500\n",
            "Epoch 87/100\n",
            "17/17 [==============================] - 2s 128ms/step - loss: 0.3480 - accuracy: 0.8963 - val_loss: 0.3850 - val_accuracy: 0.8500\n",
            "Epoch 88/100\n",
            "17/17 [==============================] - 2s 127ms/step - loss: 0.3446 - accuracy: 0.8972 - val_loss: 0.3826 - val_accuracy: 0.8500\n",
            "Epoch 89/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.3415 - accuracy: 0.8981 - val_loss: 0.3801 - val_accuracy: 0.8500\n",
            "Epoch 90/100\n",
            "17/17 [==============================] - 2s 132ms/step - loss: 0.3386 - accuracy: 0.8991 - val_loss: 0.3775 - val_accuracy: 0.8417\n",
            "Epoch 91/100\n",
            "17/17 [==============================] - 4s 218ms/step - loss: 0.3355 - accuracy: 0.8991 - val_loss: 0.3750 - val_accuracy: 0.8500\n",
            "Epoch 92/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.3327 - accuracy: 0.9000 - val_loss: 0.3725 - val_accuracy: 0.8500\n",
            "Epoch 93/100\n",
            "17/17 [==============================] - 2s 125ms/step - loss: 0.3299 - accuracy: 0.9009 - val_loss: 0.3702 - val_accuracy: 0.8583\n",
            "Epoch 94/100\n",
            "17/17 [==============================] - 2s 129ms/step - loss: 0.3269 - accuracy: 0.9009 - val_loss: 0.3672 - val_accuracy: 0.8500\n",
            "Epoch 95/100\n",
            "17/17 [==============================] - 2s 135ms/step - loss: 0.3240 - accuracy: 0.9000 - val_loss: 0.3650 - val_accuracy: 0.8583\n",
            "Epoch 96/100\n",
            "17/17 [==============================] - 4s 227ms/step - loss: 0.3212 - accuracy: 0.9009 - val_loss: 0.3626 - val_accuracy: 0.8583\n",
            "Epoch 97/100\n",
            "17/17 [==============================] - 2s 131ms/step - loss: 0.3186 - accuracy: 0.9028 - val_loss: 0.3605 - val_accuracy: 0.8500\n",
            "Epoch 98/100\n",
            "17/17 [==============================] - 2s 124ms/step - loss: 0.3161 - accuracy: 0.9028 - val_loss: 0.3579 - val_accuracy: 0.8583\n",
            "Epoch 99/100\n",
            "17/17 [==============================] - 2s 125ms/step - loss: 0.3135 - accuracy: 0.9028 - val_loss: 0.3557 - val_accuracy: 0.8583\n",
            "Epoch 100/100\n",
            "17/17 [==============================] - 3s 173ms/step - loss: 0.3109 - accuracy: 0.9037 - val_loss: 0.3535 - val_accuracy: 0.8667\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history.history"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jCkUFndrXxYN",
        "outputId": "aa5b8868-1578-460c-f9e7-afd3dad464f7"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'loss': [1.7970823049545288,\n",
              "  1.775236964225769,\n",
              "  1.7668402194976807,\n",
              "  1.7570414543151855,\n",
              "  1.7480438947677612,\n",
              "  1.735154628753662,\n",
              "  1.7141587734222412,\n",
              "  1.6877421140670776,\n",
              "  1.653903841972351,\n",
              "  1.612270474433899,\n",
              "  1.5683495998382568,\n",
              "  1.5212206840515137,\n",
              "  1.472589373588562,\n",
              "  1.4223426580429077,\n",
              "  1.3725197315216064,\n",
              "  1.326338529586792,\n",
              "  1.2831381559371948,\n",
              "  1.240975260734558,\n",
              "  1.2010729312896729,\n",
              "  1.163108468055725,\n",
              "  1.1256370544433594,\n",
              "  1.0912644863128662,\n",
              "  1.058565616607666,\n",
              "  1.0269526243209839,\n",
              "  0.9981207847595215,\n",
              "  0.9702247977256775,\n",
              "  0.9432231783866882,\n",
              "  0.9179144501686096,\n",
              "  0.8941764831542969,\n",
              "  0.8704922199249268,\n",
              "  0.8475461602210999,\n",
              "  0.8253816366195679,\n",
              "  0.80365389585495,\n",
              "  0.7834052443504333,\n",
              "  0.7640787959098816,\n",
              "  0.7452824115753174,\n",
              "  0.7273719310760498,\n",
              "  0.7099511623382568,\n",
              "  0.6933642625808716,\n",
              "  0.6776057481765747,\n",
              "  0.6627607941627502,\n",
              "  0.6481862664222717,\n",
              "  0.6346948146820068,\n",
              "  0.621802806854248,\n",
              "  0.6093968152999878,\n",
              "  0.5970624089241028,\n",
              "  0.5852219462394714,\n",
              "  0.5737836956977844,\n",
              "  0.5629658102989197,\n",
              "  0.5527868270874023,\n",
              "  0.5431178212165833,\n",
              "  0.5335296392440796,\n",
              "  0.5243914127349854,\n",
              "  0.5155674815177917,\n",
              "  0.5073316097259521,\n",
              "  0.4994295537471771,\n",
              "  0.49185696244239807,\n",
              "  0.48448407649993896,\n",
              "  0.4771393835544586,\n",
              "  0.47045713663101196,\n",
              "  0.46388185024261475,\n",
              "  0.45754772424697876,\n",
              "  0.45178720355033875,\n",
              "  0.4460602402687073,\n",
              "  0.4400646686553955,\n",
              "  0.43468254804611206,\n",
              "  0.42950233817100525,\n",
              "  0.42428839206695557,\n",
              "  0.4196920394897461,\n",
              "  0.41448840498924255,\n",
              "  0.4092824459075928,\n",
              "  0.4048166275024414,\n",
              "  0.40016523003578186,\n",
              "  0.3955400884151459,\n",
              "  0.39109835028648376,\n",
              "  0.38701263070106506,\n",
              "  0.3831890821456909,\n",
              "  0.3789006769657135,\n",
              "  0.37511834502220154,\n",
              "  0.37144115567207336,\n",
              "  0.36766517162323,\n",
              "  0.36402925848960876,\n",
              "  0.36061808466911316,\n",
              "  0.3574443757534027,\n",
              "  0.35426193475723267,\n",
              "  0.3509778082370758,\n",
              "  0.347978413105011,\n",
              "  0.34458649158477783,\n",
              "  0.34154897928237915,\n",
              "  0.3386273980140686,\n",
              "  0.3355434238910675,\n",
              "  0.3327328860759735,\n",
              "  0.32990697026252747,\n",
              "  0.32692670822143555,\n",
              "  0.32403382658958435,\n",
              "  0.3212341070175171,\n",
              "  0.31864795088768005,\n",
              "  0.3160756528377533,\n",
              "  0.3134876489639282,\n",
              "  0.3108695149421692],\n",
              " 'accuracy': [0.17777778208255768,\n",
              "  0.22685185074806213,\n",
              "  0.26944443583488464,\n",
              "  0.3157407343387604,\n",
              "  0.3194444477558136,\n",
              "  0.33796295523643494,\n",
              "  0.37037035822868347,\n",
              "  0.3888888955116272,\n",
              "  0.39537036418914795,\n",
              "  0.4157407283782959,\n",
              "  0.432407408952713,\n",
              "  0.4546296298503876,\n",
              "  0.4833333194255829,\n",
              "  0.5138888955116272,\n",
              "  0.5416666865348816,\n",
              "  0.5611110925674438,\n",
              "  0.5685185194015503,\n",
              "  0.5768518447875977,\n",
              "  0.5888888835906982,\n",
              "  0.5990740656852722,\n",
              "  0.6101852059364319,\n",
              "  0.6231481432914734,\n",
              "  0.6314814686775208,\n",
              "  0.6416666507720947,\n",
              "  0.6537036895751953,\n",
              "  0.6712962985038757,\n",
              "  0.6824073791503906,\n",
              "  0.6944444179534912,\n",
              "  0.699999988079071,\n",
              "  0.7120370268821716,\n",
              "  0.7185184955596924,\n",
              "  0.7240740656852722,\n",
              "  0.7388888597488403,\n",
              "  0.7435185313224792,\n",
              "  0.75,\n",
              "  0.7564814686775208,\n",
              "  0.7657407522201538,\n",
              "  0.7740740776062012,\n",
              "  0.7842592597007751,\n",
              "  0.7888888716697693,\n",
              "  0.7916666865348816,\n",
              "  0.7972221970558167,\n",
              "  0.8018518686294556,\n",
              "  0.8101851940155029,\n",
              "  0.8157407641410828,\n",
              "  0.8194444179534912,\n",
              "  0.8277778029441833,\n",
              "  0.8324074149131775,\n",
              "  0.8361111283302307,\n",
              "  0.8407407402992249,\n",
              "  0.8425925970077515,\n",
              "  0.8462963104248047,\n",
              "  0.8490740656852722,\n",
              "  0.8500000238418579,\n",
              "  0.854629635810852,\n",
              "  0.8537036776542664,\n",
              "  0.855555534362793,\n",
              "  0.8574073910713196,\n",
              "  0.8601852059364319,\n",
              "  0.8611111044883728,\n",
              "  0.864814817905426,\n",
              "  0.8657407164573669,\n",
              "  0.8685185313224792,\n",
              "  0.8675925731658936,\n",
              "  0.8703703880310059,\n",
              "  0.8703703880310059,\n",
              "  0.875,\n",
              "  0.8777777552604675,\n",
              "  0.8796296119689941,\n",
              "  0.8787037134170532,\n",
              "  0.8805555701255798,\n",
              "  0.8805555701255798,\n",
              "  0.8824074268341064,\n",
              "  0.885185182094574,\n",
              "  0.8833333253860474,\n",
              "  0.8833333253860474,\n",
              "  0.8842592835426331,\n",
              "  0.8870370388031006,\n",
              "  0.8861111402511597,\n",
              "  0.8898147940635681,\n",
              "  0.8907407522201538,\n",
              "  0.8925926089286804,\n",
              "  0.8925926089286804,\n",
              "  0.895370364189148,\n",
              "  0.895370364189148,\n",
              "  0.895370364189148,\n",
              "  0.8962963223457336,\n",
              "  0.8972222208976746,\n",
              "  0.8981481194496155,\n",
              "  0.8990740776062012,\n",
              "  0.8990740776062012,\n",
              "  0.8999999761581421,\n",
              "  0.9009259343147278,\n",
              "  0.9009259343147278,\n",
              "  0.8999999761581421,\n",
              "  0.9009259343147278,\n",
              "  0.9027777910232544,\n",
              "  0.9027777910232544,\n",
              "  0.9027777910232544,\n",
              "  0.9037036895751953],\n",
              " 'val_loss': [1.7872159481048584,\n",
              "  1.7804032564163208,\n",
              "  1.7736634016036987,\n",
              "  1.767261266708374,\n",
              "  1.7606539726257324,\n",
              "  1.748331069946289,\n",
              "  1.7308075428009033,\n",
              "  1.7074122428894043,\n",
              "  1.679026484489441,\n",
              "  1.6475446224212646,\n",
              "  1.6132155656814575,\n",
              "  1.575021505355835,\n",
              "  1.5377782583236694,\n",
              "  1.489348292350769,\n",
              "  1.4424301385879517,\n",
              "  1.3964728116989136,\n",
              "  1.3535548448562622,\n",
              "  1.3090540170669556,\n",
              "  1.264355182647705,\n",
              "  1.2247109413146973,\n",
              "  1.1834914684295654,\n",
              "  1.1473981142044067,\n",
              "  1.1136399507522583,\n",
              "  1.0791064500808716,\n",
              "  1.0489259958267212,\n",
              "  1.0197944641113281,\n",
              "  0.9909042119979858,\n",
              "  0.9633496403694153,\n",
              "  0.9362556338310242,\n",
              "  0.9096651673316956,\n",
              "  0.8839887976646423,\n",
              "  0.8600490689277649,\n",
              "  0.8369351029396057,\n",
              "  0.8161026239395142,\n",
              "  0.795779287815094,\n",
              "  0.7768422365188599,\n",
              "  0.7588589787483215,\n",
              "  0.7413874268531799,\n",
              "  0.7245188355445862,\n",
              "  0.7085500359535217,\n",
              "  0.6933020353317261,\n",
              "  0.6787140369415283,\n",
              "  0.6649706959724426,\n",
              "  0.6519398093223572,\n",
              "  0.638674795627594,\n",
              "  0.6270809173583984,\n",
              "  0.614990770816803,\n",
              "  0.6044818758964539,\n",
              "  0.5934851169586182,\n",
              "  0.5835046172142029,\n",
              "  0.5737521052360535,\n",
              "  0.5639711618423462,\n",
              "  0.5544849634170532,\n",
              "  0.5454436540603638,\n",
              "  0.5368980169296265,\n",
              "  0.5288192629814148,\n",
              "  0.5204555988311768,\n",
              "  0.5132303833961487,\n",
              "  0.5058803558349609,\n",
              "  0.4990493953227997,\n",
              "  0.492721825838089,\n",
              "  0.4872031509876251,\n",
              "  0.48091834783554077,\n",
              "  0.47605597972869873,\n",
              "  0.4695749282836914,\n",
              "  0.4647650718688965,\n",
              "  0.4594227373600006,\n",
              "  0.45498529076576233,\n",
              "  0.449674516916275,\n",
              "  0.44527143239974976,\n",
              "  0.44067075848579407,\n",
              "  0.43590185046195984,\n",
              "  0.4314669966697693,\n",
              "  0.42756137251853943,\n",
              "  0.4237622618675232,\n",
              "  0.42000848054885864,\n",
              "  0.416554719209671,\n",
              "  0.41352030634880066,\n",
              "  0.40989527106285095,\n",
              "  0.4069601893424988,\n",
              "  0.4033517837524414,\n",
              "  0.40029963850975037,\n",
              "  0.39704105257987976,\n",
              "  0.39382559061050415,\n",
              "  0.3905177414417267,\n",
              "  0.3881746232509613,\n",
              "  0.3850426971912384,\n",
              "  0.3825683891773224,\n",
              "  0.3801158368587494,\n",
              "  0.3775387406349182,\n",
              "  0.3749901354312897,\n",
              "  0.3724932372570038,\n",
              "  0.37019380927085876,\n",
              "  0.36721566319465637,\n",
              "  0.36499977111816406,\n",
              "  0.362625390291214,\n",
              "  0.3604598939418793,\n",
              "  0.35786545276641846,\n",
              "  0.35572415590286255,\n",
              "  0.35345950722694397],\n",
              " 'val_accuracy': [0.17499999701976776,\n",
              "  0.1666666716337204,\n",
              "  0.2666666805744171,\n",
              "  0.2750000059604645,\n",
              "  0.28333333134651184,\n",
              "  0.30000001192092896,\n",
              "  0.28333333134651184,\n",
              "  0.3083333373069763,\n",
              "  0.3166666626930237,\n",
              "  0.3083333373069763,\n",
              "  0.3166666626930237,\n",
              "  0.3333333432674408,\n",
              "  0.3333333432674408,\n",
              "  0.3916666805744171,\n",
              "  0.42500001192092896,\n",
              "  0.44999998807907104,\n",
              "  0.4749999940395355,\n",
              "  0.5,\n",
              "  0.5166666507720947,\n",
              "  0.5249999761581421,\n",
              "  0.5416666865348816,\n",
              "  0.5583333373069763,\n",
              "  0.5666666626930237,\n",
              "  0.5666666626930237,\n",
              "  0.574999988079071,\n",
              "  0.5833333134651184,\n",
              "  0.625,\n",
              "  0.625,\n",
              "  0.6416666507720947,\n",
              "  0.6666666865348816,\n",
              "  0.6583333611488342,\n",
              "  0.6666666865348816,\n",
              "  0.6916666626930237,\n",
              "  0.7083333134651184,\n",
              "  0.7416666746139526,\n",
              "  0.75,\n",
              "  0.7583333253860474,\n",
              "  0.7583333253860474,\n",
              "  0.7749999761581421,\n",
              "  0.7666666507720947,\n",
              "  0.7666666507720947,\n",
              "  0.7749999761581421,\n",
              "  0.7749999761581421,\n",
              "  0.7749999761581421,\n",
              "  0.7916666865348816,\n",
              "  0.7833333611488342,\n",
              "  0.7916666865348816,\n",
              "  0.800000011920929,\n",
              "  0.800000011920929,\n",
              "  0.800000011920929,\n",
              "  0.800000011920929,\n",
              "  0.8083333373069763,\n",
              "  0.8083333373069763,\n",
              "  0.8083333373069763,\n",
              "  0.8083333373069763,\n",
              "  0.8083333373069763,\n",
              "  0.8166666626930237,\n",
              "  0.824999988079071,\n",
              "  0.8333333134651184,\n",
              "  0.8333333134651184,\n",
              "  0.8333333134651184,\n",
              "  0.8333333134651184,\n",
              "  0.8333333134651184,\n",
              "  0.8333333134651184,\n",
              "  0.8333333134651184,\n",
              "  0.8333333134651184,\n",
              "  0.8333333134651184,\n",
              "  0.8333333134651184,\n",
              "  0.8416666388511658,\n",
              "  0.8416666388511658,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8416666388511658,\n",
              "  0.8500000238418579,\n",
              "  0.8500000238418579,\n",
              "  0.8583333492279053,\n",
              "  0.8500000238418579,\n",
              "  0.8583333492279053,\n",
              "  0.8583333492279053,\n",
              "  0.8500000238418579,\n",
              "  0.8583333492279053,\n",
              "  0.8583333492279053,\n",
              "  0.8666666746139526]}"
            ]
          },
          "metadata": {},
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# The history.history[\"loss\"] entry is a dictionary with as many values as epochs that the model was trained on. \n",
        "df_loss_acc = pd.DataFrame(history.history)\n",
        "df_loss = df_loss_acc[['loss', 'val_loss']]\n",
        "df_loss.rename(columns={'loss':'train','val_loss':'validation'},inplace=True)\n",
        "df_acc= df_loss_acc[['accuracy','val_accuracy']]\n",
        "df_acc.rename(columns={'accuracy':'train','val_accuracy':'validation'},inplace=True)\n",
        "df_loss.plot(title='Model loss',figsize=(12,8)).set(xlabel='Epoch',ylabel='Loss')\n",
        "df_acc.plot(title='Model Accuracy',figsize=(12,8)).set(xlabel='Epoch',ylabel='Accuracy')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "9YtvcvOgZodJ",
        "outputId": "c82cbd01-7295-48cc-e63f-d912b534c653"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.8/dist-packages/pandas/core/frame.py:5039: SettingWithCopyWarning: \n",
            "A value is trying to be set on a copy of a slice from a DataFrame\n",
            "\n",
            "See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy\n",
            "  return super().rename(\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[Text(0.5, 0, 'Epoch'), Text(0, 0.5, 'Accuracy')]"
            ]
          },
          "metadata": {},
          "execution_count": 21
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x576 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "CnmgtcqwaM5b"
      },
      "execution_count": 21,
      "outputs": []
    }
  ]
}