Cat_Vs_Dog_using_Transfer_Learning.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Cat_Vs_Dog using Transfer Learning.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyMJzj3WGW48m9gaNhq5ZGpj",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/Arpitkamal/Cat-Vs-Dog-with-Data-Augmentation/blob/main/Cat_Vs_Dog_using_Transfer_Learning.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "youDfnzS_tN1",
        "outputId": "8e74271b-d8dc-421b-f8bf-c9663f73dbe3"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2022-01-16 17:48:29--  https://storage.googleapis.com/mledu-datasets/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "Resolving storage.googleapis.com (storage.googleapis.com)... 108.177.12.128, 108.177.13.128, 74.125.26.128, ...\n",
            "Connecting to storage.googleapis.com (storage.googleapis.com)|108.177.12.128|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 87910968 (84M) [application/x-hdf]\n",
            "Saving to: ‘/tmp/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5’\n",
            "\n",
            "/tmp/inception_v3_w 100%[===================>]  83.84M   291MB/s    in 0.3s    \n",
            "\n",
            "2022-01-16 17:48:29 (291 MB/s) - ‘/tmp/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5’ saved [87910968/87910968]\n",
            "\n",
            "last layer output shape:  (None, 7, 7, 768)\n"
          ]
        }
      ],
      "source": [
        "import os\n",
        "\n",
        "from tensorflow.keras import layers\n",
        "from tensorflow.keras import Model\n",
        "!wget --no-check-certificate \\\n",
        "    https://storage.googleapis.com/mledu-datasets/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5 \\\n",
        "    -O /tmp/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
        "  \n",
        "from tensorflow.keras.applications.inception_v3 import InceptionV3\n",
        "\n",
        "local_weights_file = '/tmp/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5'\n",
        "\n",
        "pre_trained_model = InceptionV3(input_shape = (150, 150, 3), \n",
        "                                include_top = False, \n",
        "                                weights = None)\n",
        "\n",
        "pre_trained_model.load_weights(local_weights_file)\n",
        "\n",
        "# locked layer\n",
        "for layer in pre_trained_model.layers:\n",
        "  layer.trainable = False\n",
        "  \n",
        "#pre_trained_model.summary()\n",
        "\n",
        "last_layer = pre_trained_model.get_layer('mixed7')\n",
        "print('last layer output shape: ', last_layer.output_shape)\n",
        "last_output = last_layer.output"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.optimizers import RMSprop\n",
        "\n",
        "# Flatten the output layer to 1 dimension\n",
        "x = layers.Flatten()(last_output)\n",
        "# Add a fully connected layer with 1,024 hidden units and ReLU activation\n",
        "x = layers.Dense(1024, activation='relu')(x)\n",
        "# Add a dropout rate of 0.2\n",
        "x = layers.Dropout(0.2)(x)                  \n",
        "# Add a final sigmoid layer for classification\n",
        "x = layers.Dense  (1, activation='sigmoid')(x)           \n",
        "\n",
        "model = Model( pre_trained_model.input, x) \n",
        "\n",
        "model.compile(optimizer = RMSprop(lr=0.0001), \n",
        "              loss = 'binary_crossentropy', \n",
        "              metrics = ['accuracy'])"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-Sy5W2aC_2-G",
        "outputId": "96735806-39fc-435f-cdea-4e8ac8758928"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/optimizer_v2/rmsprop.py:130: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
            "  super(RMSprop, self).__init__(name, **kwargs)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!wget --no-check-certificate \\\n",
        "        https://storage.googleapis.com/mledu-datasets/cats_and_dogs_filtered.zip \\\n",
        "       -O /tmp/cats_and_dogs_filtered.zip\n",
        "\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "\n",
        "import os\n",
        "import zipfile\n",
        "\n",
        "local_zip = '//tmp/cats_and_dogs_filtered.zip'\n",
        "\n",
        "zip_ref = zipfile.ZipFile(local_zip, 'r')\n",
        "\n",
        "zip_ref.extractall('/tmp')\n",
        "zip_ref.close()\n",
        "\n",
        "# Define our example directories and files\n",
        "base_dir = '/tmp/cats_and_dogs_filtered'\n",
        "\n",
        "train_dir = os.path.join( base_dir, 'train')\n",
        "validation_dir = os.path.join( base_dir, 'validation')\n",
        "\n",
        "\n",
        "train_cats_dir = os.path.join(train_dir, 'cats') # Directory with our training cat pictures\n",
        "train_dogs_dir = os.path.join(train_dir, 'dogs') # Directory with our training dog pictures\n",
        "validation_cats_dir = os.path.join(validation_dir, 'cats') # Directory with our validation cat pictures\n",
        "validation_dogs_dir = os.path.join(validation_dir, 'dogs')# Directory with our validation dog pictures\n",
        "\n",
        "train_cat_fnames = os.listdir(train_cats_dir)\n",
        "train_dog_fnames = os.listdir(train_dogs_dir)\n",
        "\n",
        "# Adding data-augmentation parameters to ImageDataGenerator in training data\n",
        "train_datagen = ImageDataGenerator(rescale = 1./255.,\n",
        "                                   rotation_range = 40,\n",
        "                                   width_shift_range = 0.2,\n",
        "                                   height_shift_range = 0.2,\n",
        "                                   shear_range = 0.2,\n",
        "                                   zoom_range = 0.2,\n",
        "                                   horizontal_flip = True)\n",
        "\n",
        "# Adding data-augmentation parameters to ImageDataGenerator in testing data\n",
        "test_datagen = ImageDataGenerator( rescale = 1.0/255.,\n",
        "                                   rotation_range = 40,\n",
        "                                   width_shift_range = 0.2,\n",
        "                                   height_shift_range = 0.2,\n",
        "                                   shear_range = 0.2,\n",
        "                                   zoom_range = 0.2,\n",
        "                                   horizontal_flip = True)\n",
        "\n",
        "# Flow training images in batches of 20 using train_datagen generator\n",
        "train_generator = train_datagen.flow_from_directory(train_dir,\n",
        "                                                    batch_size = 20,\n",
        "                                                    class_mode = 'binary', \n",
        "                                                    target_size = (150, 150))     \n",
        "\n",
        "# Flow validation images in batches of 20 using test_datagen generator\n",
        "validation_generator =  test_datagen.flow_from_directory( validation_dir,\n",
        "                                                          batch_size  = 20,\n",
        "                                                          class_mode  = 'binary', \n",
        "                                                          target_size = (150, 150))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XWI1se3k_-1Q",
        "outputId": "9eff0246-a633-4e9d-ed3d-2eb0b5a2e597"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2022-01-16 17:48:33--  https://storage.googleapis.com/mledu-datasets/cats_and_dogs_filtered.zip\n",
            "Resolving storage.googleapis.com (storage.googleapis.com)... 173.194.210.128, 173.194.211.128, 173.194.212.128, ...\n",
            "Connecting to storage.googleapis.com (storage.googleapis.com)|173.194.210.128|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 68606236 (65M) [application/zip]\n",
            "Saving to: ‘/tmp/cats_and_dogs_filtered.zip’\n",
            "\n",
            "/tmp/cats_and_dogs_ 100%[===================>]  65.43M   252MB/s    in 0.3s    \n",
            "\n",
            "2022-01-16 17:48:34 (252 MB/s) - ‘/tmp/cats_and_dogs_filtered.zip’ saved [68606236/68606236]\n",
            "\n",
            "Found 2000 images belonging to 2 classes.\n",
            "Found 1000 images belonging to 2 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "history = model.fit(\n",
        "            train_generator,\n",
        "            validation_data = validation_generator,\n",
        "            steps_per_epoch = 100,\n",
        "            epochs = 50,\n",
        "            validation_steps = 50,\n",
        "            verbose = 2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "zMhFd4reAHfy",
        "outputId": "1466c703-25c1-4ca5-dac6-db480e837df1"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/50\n",
            "100/100 - 35s - loss: 0.3423 - accuracy: 0.8720 - val_loss: 0.2076 - val_accuracy: 0.9190 - 35s/epoch - 347ms/step\n",
            "Epoch 2/50\n",
            "100/100 - 22s - loss: 0.2109 - accuracy: 0.9140 - val_loss: 0.1780 - val_accuracy: 0.9270 - 22s/epoch - 221ms/step\n",
            "Epoch 3/50\n",
            "100/100 - 22s - loss: 0.2271 - accuracy: 0.9190 - val_loss: 0.1692 - val_accuracy: 0.9400 - 22s/epoch - 218ms/step\n",
            "Epoch 4/50\n",
            "100/100 - 22s - loss: 0.2029 - accuracy: 0.9220 - val_loss: 0.2100 - val_accuracy: 0.9320 - 22s/epoch - 218ms/step\n",
            "Epoch 5/50\n",
            "100/100 - 22s - loss: 0.1796 - accuracy: 0.9410 - val_loss: 0.2101 - val_accuracy: 0.9270 - 22s/epoch - 219ms/step\n",
            "Epoch 6/50\n",
            "100/100 - 22s - loss: 0.1911 - accuracy: 0.9340 - val_loss: 0.1870 - val_accuracy: 0.9350 - 22s/epoch - 218ms/step\n",
            "Epoch 7/50\n",
            "100/100 - 22s - loss: 0.1659 - accuracy: 0.9470 - val_loss: 0.2330 - val_accuracy: 0.9260 - 22s/epoch - 219ms/step\n",
            "Epoch 8/50\n",
            "100/100 - 22s - loss: 0.1805 - accuracy: 0.9445 - val_loss: 0.2545 - val_accuracy: 0.9160 - 22s/epoch - 219ms/step\n",
            "Epoch 9/50\n",
            "100/100 - 22s - loss: 0.1582 - accuracy: 0.9395 - val_loss: 0.2587 - val_accuracy: 0.9220 - 22s/epoch - 218ms/step\n",
            "Epoch 10/50\n",
            "100/100 - 22s - loss: 0.1532 - accuracy: 0.9425 - val_loss: 0.1735 - val_accuracy: 0.9450 - 22s/epoch - 220ms/step\n",
            "Epoch 11/50\n",
            "100/100 - 22s - loss: 0.1739 - accuracy: 0.9450 - val_loss: 0.2050 - val_accuracy: 0.9410 - 22s/epoch - 218ms/step\n",
            "Epoch 12/50\n",
            "100/100 - 22s - loss: 0.1185 - accuracy: 0.9575 - val_loss: 0.1997 - val_accuracy: 0.9380 - 22s/epoch - 221ms/step\n",
            "Epoch 13/50\n",
            "100/100 - 22s - loss: 0.1518 - accuracy: 0.9490 - val_loss: 0.2236 - val_accuracy: 0.9300 - 22s/epoch - 218ms/step\n",
            "Epoch 14/50\n",
            "100/100 - 22s - loss: 0.1668 - accuracy: 0.9465 - val_loss: 0.1776 - val_accuracy: 0.9370 - 22s/epoch - 218ms/step\n",
            "Epoch 15/50\n",
            "100/100 - 22s - loss: 0.1561 - accuracy: 0.9515 - val_loss: 0.2429 - val_accuracy: 0.9310 - 22s/epoch - 218ms/step\n",
            "Epoch 16/50\n",
            "100/100 - 22s - loss: 0.1559 - accuracy: 0.9475 - val_loss: 0.2368 - val_accuracy: 0.9320 - 22s/epoch - 218ms/step\n",
            "Epoch 17/50\n",
            "100/100 - 22s - loss: 0.1381 - accuracy: 0.9560 - val_loss: 0.2490 - val_accuracy: 0.9340 - 22s/epoch - 217ms/step\n",
            "Epoch 18/50\n",
            "100/100 - 22s - loss: 0.1348 - accuracy: 0.9540 - val_loss: 0.1820 - val_accuracy: 0.9380 - 22s/epoch - 217ms/step\n",
            "Epoch 19/50\n",
            "100/100 - 22s - loss: 0.1205 - accuracy: 0.9515 - val_loss: 0.2312 - val_accuracy: 0.9380 - 22s/epoch - 216ms/step\n",
            "Epoch 20/50\n",
            "100/100 - 22s - loss: 0.1188 - accuracy: 0.9585 - val_loss: 0.2368 - val_accuracy: 0.9330 - 22s/epoch - 217ms/step\n",
            "Epoch 21/50\n",
            "100/100 - 22s - loss: 0.1392 - accuracy: 0.9525 - val_loss: 0.1942 - val_accuracy: 0.9330 - 22s/epoch - 218ms/step\n",
            "Epoch 22/50\n",
            "100/100 - 22s - loss: 0.1220 - accuracy: 0.9605 - val_loss: 0.2440 - val_accuracy: 0.9300 - 22s/epoch - 216ms/step\n",
            "Epoch 23/50\n",
            "100/100 - 22s - loss: 0.1280 - accuracy: 0.9570 - val_loss: 0.2773 - val_accuracy: 0.9260 - 22s/epoch - 217ms/step\n",
            "Epoch 24/50\n",
            "100/100 - 22s - loss: 0.1230 - accuracy: 0.9605 - val_loss: 0.2073 - val_accuracy: 0.9300 - 22s/epoch - 216ms/step\n",
            "Epoch 25/50\n",
            "100/100 - 22s - loss: 0.1085 - accuracy: 0.9660 - val_loss: 0.2185 - val_accuracy: 0.9420 - 22s/epoch - 217ms/step\n",
            "Epoch 26/50\n",
            "100/100 - 22s - loss: 0.0977 - accuracy: 0.9635 - val_loss: 0.2462 - val_accuracy: 0.9420 - 22s/epoch - 217ms/step\n",
            "Epoch 27/50\n",
            "100/100 - 22s - loss: 0.1285 - accuracy: 0.9610 - val_loss: 0.2150 - val_accuracy: 0.9390 - 22s/epoch - 218ms/step\n",
            "Epoch 28/50\n",
            "100/100 - 22s - loss: 0.0975 - accuracy: 0.9650 - val_loss: 0.2237 - val_accuracy: 0.9340 - 22s/epoch - 218ms/step\n",
            "Epoch 29/50\n",
            "100/100 - 22s - loss: 0.1077 - accuracy: 0.9665 - val_loss: 0.1800 - val_accuracy: 0.9480 - 22s/epoch - 217ms/step\n",
            "Epoch 30/50\n",
            "100/100 - 22s - loss: 0.1214 - accuracy: 0.9645 - val_loss: 0.2189 - val_accuracy: 0.9440 - 22s/epoch - 217ms/step\n",
            "Epoch 31/50\n",
            "100/100 - 22s - loss: 0.1050 - accuracy: 0.9640 - val_loss: 0.2350 - val_accuracy: 0.9390 - 22s/epoch - 218ms/step\n",
            "Epoch 32/50\n",
            "100/100 - 22s - loss: 0.1109 - accuracy: 0.9630 - val_loss: 0.2151 - val_accuracy: 0.9440 - 22s/epoch - 217ms/step\n",
            "Epoch 33/50\n",
            "100/100 - 22s - loss: 0.1022 - accuracy: 0.9610 - val_loss: 0.2104 - val_accuracy: 0.9430 - 22s/epoch - 216ms/step\n",
            "Epoch 34/50\n",
            "100/100 - 22s - loss: 0.1064 - accuracy: 0.9640 - val_loss: 0.2302 - val_accuracy: 0.9420 - 22s/epoch - 218ms/step\n",
            "Epoch 35/50\n",
            "100/100 - 22s - loss: 0.1045 - accuracy: 0.9675 - val_loss: 0.2341 - val_accuracy: 0.9410 - 22s/epoch - 216ms/step\n",
            "Epoch 36/50\n",
            "100/100 - 22s - loss: 0.0917 - accuracy: 0.9710 - val_loss: 0.3125 - val_accuracy: 0.9350 - 22s/epoch - 217ms/step\n",
            "Epoch 37/50\n",
            "100/100 - 22s - loss: 0.0844 - accuracy: 0.9720 - val_loss: 0.2296 - val_accuracy: 0.9380 - 22s/epoch - 217ms/step\n",
            "Epoch 38/50\n",
            "100/100 - 22s - loss: 0.0943 - accuracy: 0.9715 - val_loss: 0.2252 - val_accuracy: 0.9440 - 22s/epoch - 216ms/step\n",
            "Epoch 39/50\n",
            "100/100 - 22s - loss: 0.0809 - accuracy: 0.9770 - val_loss: 0.2419 - val_accuracy: 0.9410 - 22s/epoch - 216ms/step\n",
            "Epoch 40/50\n",
            "100/100 - 22s - loss: 0.1001 - accuracy: 0.9680 - val_loss: 0.2213 - val_accuracy: 0.9390 - 22s/epoch - 218ms/step\n",
            "Epoch 41/50\n",
            "100/100 - 22s - loss: 0.1041 - accuracy: 0.9665 - val_loss: 0.2456 - val_accuracy: 0.9370 - 22s/epoch - 217ms/step\n",
            "Epoch 42/50\n",
            "100/100 - 22s - loss: 0.0935 - accuracy: 0.9700 - val_loss: 0.2132 - val_accuracy: 0.9440 - 22s/epoch - 218ms/step\n",
            "Epoch 43/50\n",
            "100/100 - 22s - loss: 0.0872 - accuracy: 0.9660 - val_loss: 0.2335 - val_accuracy: 0.9300 - 22s/epoch - 218ms/step\n",
            "Epoch 44/50\n",
            "100/100 - 22s - loss: 0.0737 - accuracy: 0.9745 - val_loss: 0.2341 - val_accuracy: 0.9390 - 22s/epoch - 216ms/step\n",
            "Epoch 45/50\n",
            "100/100 - 22s - loss: 0.1062 - accuracy: 0.9690 - val_loss: 0.2059 - val_accuracy: 0.9370 - 22s/epoch - 218ms/step\n",
            "Epoch 46/50\n",
            "100/100 - 22s - loss: 0.0885 - accuracy: 0.9710 - val_loss: 0.2665 - val_accuracy: 0.9380 - 22s/epoch - 217ms/step\n",
            "Epoch 47/50\n",
            "100/100 - 22s - loss: 0.0823 - accuracy: 0.9745 - val_loss: 0.1625 - val_accuracy: 0.9530 - 22s/epoch - 219ms/step\n",
            "Epoch 48/50\n",
            "100/100 - 22s - loss: 0.0964 - accuracy: 0.9690 - val_loss: 0.2270 - val_accuracy: 0.9390 - 22s/epoch - 221ms/step\n",
            "Epoch 49/50\n",
            "100/100 - 22s - loss: 0.0842 - accuracy: 0.9735 - val_loss: 0.2026 - val_accuracy: 0.9470 - 22s/epoch - 219ms/step\n",
            "Epoch 50/50\n",
            "100/100 - 22s - loss: 0.0673 - accuracy: 0.9775 - val_loss: 0.2918 - val_accuracy: 0.9350 - 22s/epoch - 217ms/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt\n",
        "acc = history.history['accuracy']\n",
        "val_acc = history.history['val_accuracy']\n",
        "loss = history.history['loss']\n",
        "val_loss = history.history['val_loss']\n",
        "\n",
        "epochs = range(len(acc))\n",
        "\n",
        "plt.plot(epochs, acc, 'r', label='Training accuracy')\n",
        "plt.plot(epochs, val_acc, 'b', label='Validation accuracy')\n",
        "plt.title('Training and validation accuracy')\n",
        "plt.legend(loc=0)\n",
        "plt.figure()\n",
        "\n",
        "\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 300
        },
        "id": "Yj76JlivALsq",
        "outputId": "de63c021-f379-4ac3-97d4-9247426c810a"
      },
      "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": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "file_dir = '/content/Cat_dog_transfer.h5'\n",
        "model.save(file_dir)"
      ],
      "metadata": {
        "id": "N37GKlpoFY9g"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "0x6euGsWFpPA"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}