tutorial103_DL_terminology_data_augmentation2.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "tutorial103_DL_terminology_data_augmentation2.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "mount_file_id": "1mr1ThepWeXzPbQqJtAB6d6aKtFXlrE5z",
      "authorship_tag": "ABX9TyOoZKuDqiUOJUCNgSoTZVTC",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/bnsreenu/python_for_image_processing_APEER/blob/master/tutorial103_DL_terminology_data_augmentation2.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vw3JWa6yJSHi"
      },
      "source": [
        "https://youtu.be/_bahDWWufCc"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Qe8uhqkRqm-B"
      },
      "source": [
        "**Understanding deep learning terminology**\n",
        "\n",
        "This code can be used to explain data augmentation.\n",
        "\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QmFuGFkckbHL"
      },
      "source": [
        "**cifar10 dataset**\n",
        "\n",
        "60,000 32×32 pixel images divided into 10 classes.\n",
        "\n",
        "0: airplane\n",
        "\n",
        "1: automobile\n",
        "\n",
        "2: bird\n",
        "\n",
        "3: cat\n",
        "\n",
        "4: deer\n",
        "\n",
        "5: dog\n",
        "\n",
        "6: frog\n",
        "\n",
        "7: horse\n",
        "\n",
        "8: ship\n",
        "\n",
        "9: truck"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "shy9DkxGlLPK"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.image as mpimg\n",
        "import numpy as np\n",
        "\n",
        "plt.style.use('classic')\n",
        "\n",
        "from keras.models import Sequential\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras.layers import BatchNormalization\n",
        "from keras.layers import Conv2D, MaxPooling2D, Dense, Flatten\n",
        "from keras.datasets import cifar10\n",
        "from keras.utils import normalize, to_categorical\n",
        "from keras.layers import Dropout\n",
        "from keras.optimizers import SGD, RMSprop\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PhVZEd08lQAk",
        "outputId": "635cf9de-e705-4918-c4a5-d48ba29cbdef"
      },
      "source": [
        "# Input data\n",
        "\n",
        "(X_train, y_train), (X_test, y_test) = cifar10.load_data()\n",
        "print(\"The size of training dataset X is: \", X_train.shape)\n",
        "print(\"The size of testing dataset X is: \", X_test.shape)\n",
        "\n",
        "print(\"The size of training dataset y is: \", y_train.shape)\n",
        "print(\"The size of testing dataset y is: \", y_test.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The size of training dataset X is:  (50000, 32, 32, 3)\n",
            "The size of testing dataset X is:  (10000, 32, 32, 3)\n",
            "The size of training dataset y is:  (50000, 1)\n",
            "The size of testing dataset y is:  (10000, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Eryk-VzFvXbi",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "00cae234-fc33-46ba-a6fc-7ce14c48fcdf"
      },
      "source": [
        "#Decrease the dataset size to see the effect - decrease it to 1000 (test_size=0.1)\n",
        "from sklearn.model_selection import train_test_split\n",
        "_, X, _, Y = train_test_split(X_train, y_train, test_size = 0.2, random_state = 0)\n",
        "print(\"The size of the dataset X is: \", X.shape)\n",
        "print(\"The size of the dataset Y is: \", Y.shape)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The size of the dataset X is:  (10000, 32, 32, 3)\n",
            "The size of the dataset Y is:  (10000, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kHfK8axvvktR",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "57f9803f-19b4-402c-9e23-0d9c38c46978"
      },
      "source": [
        "#Split again into train and test to create small training and testing dataset\n",
        "X_train, X_test, y_train, y_test = train_test_split(X, Y, test_size = 0.25, random_state = 0)\n",
        "print(\"The size of training dataset is: \", X_train.shape)\n",
        "print(\"The size of testing dataset is: \", X_test.shape)\n",
        "print(\"The size of training dataset y is: \", y_train.shape)\n",
        "print(\"The size of testing dataset y is: \", y_test.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The size of training dataset is:  (7500, 32, 32, 3)\n",
            "The size of testing dataset is:  (2500, 32, 32, 3)\n",
            "The size of training dataset y is:  (7500, 1)\n",
            "The size of testing dataset y is:  (2500, 1)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 272
        },
        "id": "YxlG2FPPmJWN",
        "outputId": "ffce3457-f243-485f-d3a3-1a5c429a4cf2"
      },
      "source": [
        "#view few images \n",
        "for i in range(9):\n",
        "\tplt.subplot(330 + 1 + i)\n",
        "\tplt.imshow(X_train[i])\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EtvFqkillTi1"
      },
      "source": [
        "X_train = (X_train.astype('float32')) / 255.\n",
        "X_test = (X_test.astype('float32')) / 255.\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GCFK8L81YsKi",
        "outputId": "4f7c6629-6d01-4f04-bcfb-2aa96e727c29"
      },
      "source": [
        "# Print a few y_train values to see before and after categorical\n",
        "print(y_train[0])\n",
        "print(y_train[1])\n",
        "print(y_train[10])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[3]\n",
            "[1]\n",
            "[5]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lFxncFsalV5E"
      },
      "source": [
        "y_train = to_categorical(y_train)\n",
        "y_test = to_categorical(y_test)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qYq15gVCYetG",
        "outputId": "2ed9ba6d-2d64-4ed8-d014-e3836aa9277b"
      },
      "source": [
        "print(y_train[0])\n",
        "print(y_train[1])\n",
        "print(y_train[10])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[1. 0. 0. 0. 0. 0. 0. 0. 0. 0.]\n",
            "[0. 0. 0. 0. 1. 0. 0. 0. 0. 0.]\n",
            "[0. 0. 0. 0. 0. 0. 0. 1. 0. 0.]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "s2yZ1ytHUf-F",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6e362fa5-5c05-4fd8-dd82-f5349b6cbfb8"
      },
      "source": [
        "# Create a model with dropout\n",
        "drop=0.25\n",
        "\n",
        "kernel_initializer =  'he_uniform'  #Also try 'zeros', 'random_uniform', 'he_uniform', 'glorot_uniform'\n",
        "\n",
        "model1 = Sequential()\n",
        "model1.add(Conv2D(32, (3, 3), activation='relu', kernel_initializer=kernel_initializer, padding='same', input_shape=(32, 32, 3)))\n",
        "model1.add(Conv2D(32, (3, 3), activation='relu', kernel_initializer=kernel_initializer, padding='same'))\n",
        "model1.add(MaxPooling2D((2, 2)))\n",
        "model1.add(Dropout(drop))\n",
        "\n",
        "model1.add(Conv2D(64, (3, 3), activation='relu', kernel_initializer=kernel_initializer, padding='same'))\n",
        "model1.add(Conv2D(64, (3, 3), activation='relu', kernel_initializer=kernel_initializer, padding='same'))\n",
        "model1.add(MaxPooling2D((2, 2)))\n",
        "model1.add(Dropout(drop))\n",
        "\n",
        "model1.add(Flatten())\n",
        "model1.add(Dense(512, activation='relu', kernel_initializer=kernel_initializer))\n",
        "model1.add(Dropout(drop))\n",
        "model1.add(Dense(10, activation='softmax'))\n",
        "\n",
        "opt1 = SGD(lr=0.001, momentum=0.9)\n",
        "opt2 = RMSprop(lr=0.001, decay=1e-6)\n",
        "model1.compile(optimizer=opt1, loss='categorical_crossentropy', metrics=['accuracy'])\n",
        "model1.summary()\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 32, 32, 32)        896       \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 32, 32, 32)        9248      \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 16, 16, 32)        0         \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 16, 16, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 16, 16, 64)        18496     \n",
            "_________________________________________________________________\n",
            "conv2d_3 (Conv2D)            (None, 16, 16, 64)        36928     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 8, 8, 64)          0         \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 8, 8, 64)          0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 4096)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 512)               2097664   \n",
            "_________________________________________________________________\n",
            "dropout_2 (Dropout)          (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 10)                5130      \n",
            "=================================================================\n",
            "Total params: 2,168,362\n",
            "Trainable params: 2,168,362\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7SGbACGslrpb"
      },
      "source": [
        "#########################################################\n",
        "#Fit model without augmentation ....\n",
        "\n",
        "#history = model1.fit(X_train, y_train, epochs=50, batch_size=64, validation_data=(X_test, y_test), verbose=1)\n",
        "\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XlYg3MU3lrvt"
      },
      "source": [
        "###################################################\n",
        "######### Data augmentation to improve the model\n",
        "\n",
        "train_datagen = ImageDataGenerator(rotation_range=15,  #Too much rotation may hurt accuracy, especially for small datasets.\n",
        "    width_shift_range=0.1,\n",
        "    height_shift_range=0.1,\n",
        "    zoom_range = 0.1,\n",
        "    vertical_flip=False,\n",
        "    horizontal_flip = True,\n",
        "    fill_mode=\"reflect\")\n",
        "\n",
        "#train_datagen.fit(X_train)\n",
        "\n",
        "train_generator = train_datagen.flow(\n",
        "    X_train,\n",
        "    y_train,\n",
        "    batch_size = 32)  #images to generate in a batch\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8NZw_9w3xeMe",
        "outputId": "861cc85e-2bf0-4ddc-c782-cea04c766996"
      },
      "source": [
        "x = train_generator.next()\n",
        "print(x[0].shape)  #Images\n",
        "print(x[1].shape)  #Labels\n",
        "print((x[0].shape[0]))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(32, 32, 32, 3)\n",
            "(32, 10)\n",
            "32\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 137
        },
        "id": "sSTnesybxQXG",
        "outputId": "387f4862-1264-41d0-9422-dcc221dbbf55"
      },
      "source": [
        "x = train_generator.next()\n",
        "image = x[0][0]\n",
        "title = np.argmax(x[1][0])\n",
        "plt.figure(figsize=(1.5, 1.5))\n",
        "plt.suptitle(title, fontsize=12)\n",
        "plt.imshow(image)\n",
        "plt.show()\n",
        "#0: airplane , 1: automobile, 2: bird, 3: cat, 4: deer, 5: dog, 6: frog, 7: horse, 8: ship, 9: truck"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 108x108 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sXwAXikMELN7",
        "outputId": "05fe20f7-e91f-4ce7-f62d-5c3f3e37837a"
      },
      "source": [
        "print(\"Total number of training images in the dataset = \", X_train.shape[0])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Total number of training images in the dataset =  7500\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2ewjeqgcwNu8",
        "outputId": "a0ee8175-a259-46d4-8c76-9bc7d1488a2d"
      },
      "source": [
        "#NOTE: When we use fit_generator, the number of samples processed \n",
        "#for each epoch is batch_size * steps_per_epochs. \n",
        "#should typically be equal to the number of unique samples in our \n",
        "#dataset divided by the batch size.\n",
        "\n",
        "batch_size = 32   #Match this to the batch_size from generator\n",
        "steps_per_epoch = len(X_train) // batch_size  \n",
        "\n",
        "print(\"Steps per epoch = \", steps_per_epoch)\n",
        "print(\"Total data per epoch = \", steps_per_epoch*batch_size)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Steps per epoch =  234\n",
            "Total data per epoch =  7488\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zmQ3Dat7vjlq",
        "outputId": "7dbd6770-0b39-48c0-84c3-e9b7030f2f40"
      },
      "source": [
        "\n",
        "#Train using model.fit (older versions of keras, use model.fit_generator)\n",
        "\"\"\"\n",
        "history1 = model1.fit(\n",
        "        train_generator,\n",
        "        steps_per_epoch = steps_per_epoch,\n",
        "        epochs = 50,\n",
        "        validation_data = validation_generator,\n",
        "        validation_steps=val_steps_per_epoch)\n",
        "\"\"\"\n",
        "\n",
        "history = model1.fit(\n",
        "        train_generator,\n",
        "        steps_per_epoch = steps_per_epoch,\n",
        "        epochs = 50,\n",
        "        validation_data = (X_test, y_test))\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "234/234 [==============================] - 8s 20ms/step - loss: 2.4036 - accuracy: 0.1270 - val_loss: 2.0274 - val_accuracy: 0.2620\n",
            "Epoch 2/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 2.0313 - accuracy: 0.2460 - val_loss: 1.9080 - val_accuracy: 0.3264\n",
            "Epoch 3/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.9150 - accuracy: 0.3043 - val_loss: 1.7569 - val_accuracy: 0.3652\n",
            "Epoch 4/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.8216 - accuracy: 0.3280 - val_loss: 1.7305 - val_accuracy: 0.3556\n",
            "Epoch 5/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.7524 - accuracy: 0.3564 - val_loss: 1.6120 - val_accuracy: 0.4128\n",
            "Epoch 6/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.6882 - accuracy: 0.3783 - val_loss: 1.6090 - val_accuracy: 0.4140\n",
            "Epoch 7/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.6792 - accuracy: 0.3788 - val_loss: 1.6055 - val_accuracy: 0.4200\n",
            "Epoch 8/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.6228 - accuracy: 0.3979 - val_loss: 1.5570 - val_accuracy: 0.4504\n",
            "Epoch 9/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.5850 - accuracy: 0.4224 - val_loss: 1.5473 - val_accuracy: 0.4392\n",
            "Epoch 10/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.5657 - accuracy: 0.4196 - val_loss: 1.4987 - val_accuracy: 0.4616\n",
            "Epoch 11/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.5439 - accuracy: 0.4394 - val_loss: 1.4566 - val_accuracy: 0.4876\n",
            "Epoch 12/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.5116 - accuracy: 0.4461 - val_loss: 1.4163 - val_accuracy: 0.4808\n",
            "Epoch 13/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.5175 - accuracy: 0.4456 - val_loss: 1.4633 - val_accuracy: 0.4872\n",
            "Epoch 14/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.4774 - accuracy: 0.4550 - val_loss: 1.4763 - val_accuracy: 0.4800\n",
            "Epoch 15/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.4692 - accuracy: 0.4517 - val_loss: 1.3633 - val_accuracy: 0.5116\n",
            "Epoch 16/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.4325 - accuracy: 0.4817 - val_loss: 1.3727 - val_accuracy: 0.5144\n",
            "Epoch 17/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.4412 - accuracy: 0.4794 - val_loss: 1.3421 - val_accuracy: 0.5188\n",
            "Epoch 18/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.4110 - accuracy: 0.4813 - val_loss: 1.3321 - val_accuracy: 0.5208\n",
            "Epoch 19/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.3997 - accuracy: 0.4855 - val_loss: 1.3334 - val_accuracy: 0.5168\n",
            "Epoch 20/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.3943 - accuracy: 0.4793 - val_loss: 1.3144 - val_accuracy: 0.5276\n",
            "Epoch 21/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.3635 - accuracy: 0.5105 - val_loss: 1.2815 - val_accuracy: 0.5364\n",
            "Epoch 22/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.3462 - accuracy: 0.5125 - val_loss: 1.3129 - val_accuracy: 0.5228\n",
            "Epoch 23/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.3508 - accuracy: 0.5099 - val_loss: 1.2990 - val_accuracy: 0.5400\n",
            "Epoch 24/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.3368 - accuracy: 0.5158 - val_loss: 1.2585 - val_accuracy: 0.5436\n",
            "Epoch 25/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.3051 - accuracy: 0.5294 - val_loss: 1.2409 - val_accuracy: 0.5576\n",
            "Epoch 26/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.3133 - accuracy: 0.5209 - val_loss: 1.2765 - val_accuracy: 0.5508\n",
            "Epoch 27/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.2973 - accuracy: 0.5203 - val_loss: 1.2220 - val_accuracy: 0.5652\n",
            "Epoch 28/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.2866 - accuracy: 0.5366 - val_loss: 1.2635 - val_accuracy: 0.5528\n",
            "Epoch 29/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.2918 - accuracy: 0.5318 - val_loss: 1.2647 - val_accuracy: 0.5416\n",
            "Epoch 30/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.2967 - accuracy: 0.5279 - val_loss: 1.2473 - val_accuracy: 0.5592\n",
            "Epoch 31/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.2749 - accuracy: 0.5281 - val_loss: 1.2504 - val_accuracy: 0.5416\n",
            "Epoch 32/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.2638 - accuracy: 0.5449 - val_loss: 1.1884 - val_accuracy: 0.5728\n",
            "Epoch 33/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.2265 - accuracy: 0.5467 - val_loss: 1.1587 - val_accuracy: 0.5888\n",
            "Epoch 34/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.2342 - accuracy: 0.5525 - val_loss: 1.1639 - val_accuracy: 0.5724\n",
            "Epoch 35/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.2020 - accuracy: 0.5744 - val_loss: 1.1570 - val_accuracy: 0.5856\n",
            "Epoch 36/50\n",
            "234/234 [==============================] - 4s 18ms/step - loss: 1.2125 - accuracy: 0.5635 - val_loss: 1.1814 - val_accuracy: 0.5756\n",
            "Epoch 37/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.1971 - accuracy: 0.5710 - val_loss: 1.1317 - val_accuracy: 0.5956\n",
            "Epoch 38/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.2035 - accuracy: 0.5681 - val_loss: 1.2987 - val_accuracy: 0.5460\n",
            "Epoch 39/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.1873 - accuracy: 0.5729 - val_loss: 1.1929 - val_accuracy: 0.5804\n",
            "Epoch 40/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.1595 - accuracy: 0.5782 - val_loss: 1.1680 - val_accuracy: 0.5820\n",
            "Epoch 41/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.1712 - accuracy: 0.5789 - val_loss: 1.1702 - val_accuracy: 0.5876\n",
            "Epoch 42/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.1368 - accuracy: 0.5798 - val_loss: 1.1538 - val_accuracy: 0.5852\n",
            "Epoch 43/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.1548 - accuracy: 0.5860 - val_loss: 1.1320 - val_accuracy: 0.5960\n",
            "Epoch 44/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.1104 - accuracy: 0.5984 - val_loss: 1.1246 - val_accuracy: 0.5988\n",
            "Epoch 45/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.1160 - accuracy: 0.5992 - val_loss: 1.1477 - val_accuracy: 0.6020\n",
            "Epoch 46/50\n",
            "234/234 [==============================] - 5s 20ms/step - loss: 1.1187 - accuracy: 0.5959 - val_loss: 1.1180 - val_accuracy: 0.6040\n",
            "Epoch 47/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.0798 - accuracy: 0.6053 - val_loss: 1.1552 - val_accuracy: 0.6012\n",
            "Epoch 48/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.0945 - accuracy: 0.6100 - val_loss: 1.1416 - val_accuracy: 0.6020\n",
            "Epoch 49/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.1101 - accuracy: 0.5981 - val_loss: 1.0700 - val_accuracy: 0.6220\n",
            "Epoch 50/50\n",
            "234/234 [==============================] - 4s 19ms/step - loss: 1.0651 - accuracy: 0.6252 - val_loss: 1.0891 - val_accuracy: 0.6200\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "EN5DsxdVmwnt",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "1942aa70-4e93-4f60-c7d4-d775cef502ba"
      },
      "source": [
        "_, acc = model1.evaluate(X_test, y_test)\n",
        "print(\"Accuracy = \", (acc * 100.0), \"%\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "79/79 [==============================] - 0s 3ms/step - loss: 1.0891 - accuracy: 0.6200\n",
            "Accuracy =  62.00000047683716 %\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 425
        },
        "id": "NYYnwQf8r3Wr",
        "outputId": "d120b7ef-c7e8-4839-f464-12f304e4f6af"
      },
      "source": [
        "#Accuracy with and without Augmentation\n",
        "import pandas as pd\n",
        "without_aug = {1000:36.4, 2000:45.2, 5000:51.7, 10000:58.4, 25000:69.4, 50000:77.3}\n",
        "with_aug = {1000:44, 2000:48.4, 5000:54.7, 10000:60.8, 25000:70.7, 50000:78.4}\n",
        "df = pd.DataFrame([without_aug, with_aug])\n",
        "df = df.T\n",
        "df.reset_index(inplace=True)\n",
        "\n",
        "\n",
        "df.columns =['num_images', 'without_aug', 'with_aug']\n",
        "print(df.head)\n",
        "\n",
        "df.plot(x='num_images', y=['without_aug', 'with_aug'], kind='line')\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "<bound method NDFrame.head of    num_images  without_aug  with_aug\n",
            "0        1000         36.4      44.0\n",
            "1        2000         45.2      48.4\n",
            "2        5000         51.7      54.7\n",
            "3       10000         58.4      60.8\n",
            "4       25000         69.4      70.7\n",
            "5       50000         77.3      78.4>\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7f225c0e06d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 34
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E2dhVAMfUBO6"
      },
      "source": [
        "Number of images correspond to the toal images including training and validation. \n",
        "\n",
        "\n",
        "Without augmentation and 50 epochs:\n",
        "\n",
        "- Accuracy with 1000 images: 36.4%\n",
        "- Accuracy with 2000 images: 45.2%\n",
        "- Accuracy with 5000 images: 51.7%\n",
        "- Accuracy with 10000 images: 58.4%\n",
        "- Accuracy with 25000 images: 69.4%\n",
        "- Accuracy with 50000 images: 77.3%\n",
        "\n",
        "With augmentation and 50 epochs:\n",
        "\n",
        "- Accuracy with 1000 images: 44.0%\n",
        "- Accuracy with 2000 images: 48.4%\n",
        "- Accuracy with 5000 images: 54.7%\n",
        "- Accuracy with 10000 images: 60.8%\n",
        "- Accuracy with 25000 images: 70.7%\n",
        "- Accuracy with 50000 images: 78.4%\n",
        "\n",
        "If your accuracy from augmentation is worse compared to without augmentation, it means your augmentation is notworking out for you. You need to tweak it, for example by limiting rotation or changing fill mode, etc. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ex9OhzR8XS4d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 579
        },
        "outputId": "695ac2a6-5431-4139-943f-d8ced9b839e3"
      },
      "source": [
        "#####################################################################\n",
        "#plot the training and validation accuracy and loss at each epoch\n",
        "#If validation loss is lower than training loss this could be becuase we are applying\n",
        "#regularization (Dropout) during training which won't be applied during validation. \n",
        "#Also, training loss is measured during each epoch while validation is done after the epoch. \n",
        "\n",
        "history = history\n",
        "\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "epochs = range(1, len(loss) + 1)\n",
        "plt.plot(epochs, loss, 'y', label='Training loss')\n",
        "plt.plot(epochs, val_loss, 'r', label='Validation loss')\n",
        "plt.title('Training and validation loss')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Loss')\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "\n",
        "acc = history.history['accuracy']\n",
        "val_acc = history.history['val_accuracy']\n",
        "plt.plot(epochs, acc, 'y', label='Training acc')\n",
        "plt.plot(epochs, val_acc, 'r', label='Validation acc')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.xlabel('Epochs')\n",
        "plt.ylabel('Accuracy')\n",
        "plt.legend()\n",
        "plt.show()\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}