Transfer_learning_with_inception_v3.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Transfer learning with inception v3",
      "provenance": [],
      "authorship_tag": "ABX9TyMFF98mYCzjr6KVvTour4bD",
      "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/udaysrinu/Tensorflow-classification-models/blob/master/Transfer_learning_with_inception_v3.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lp4_GB1VQwQ1",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 218
        },
        "outputId": "e3f8d45d-8387-4f3b-b1fc-725ece6985d3"
      },
      "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",
        "for layer in pre_trained_model.layers:\n",
        "  layer.trainable = False\n",
        "  \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"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2020-08-29 03:35:56--  https://storage.googleapis.com/mledu-datasets/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "Resolving storage.googleapis.com (storage.googleapis.com)... 74.125.24.128, 74.125.200.128, 74.125.130.128, ...\n",
            "Connecting to storage.googleapis.com (storage.googleapis.com)|74.125.24.128|: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  39.2MB/s    in 2.1s    \n",
            "\n",
            "2020-08-29 03:35:59 (39.2 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"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5FuiTPIvQxa6",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        "from tensorflow.keras.optimizers import RMSprop\n",
        "\n",
        "x = layers.Flatten()(last_output)\n",
        "x = layers.Dense(1024, activation='relu')(x)\n",
        "x = layers.Dropout(0.2)(x)                  \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'])"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xFuCVCs6Q5Ec",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 235
        },
        "outputId": "0a67d1d1-bb4a-40d0-8afc-2a5f044e1169"
      },
      "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",
        "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') \n",
        "train_dogs_dir = os.path.join(train_dir, 'dogs')\n",
        "validation_cats_dir = os.path.join(validation_dir, 'cats') \n",
        "validation_dogs_dir = os.path.join(validation_dir, 'dogs')\n",
        "\n",
        "train_cat_fnames = os.listdir(train_cats_dir)\n",
        "train_dog_fnames = os.listdir(train_dogs_dir)\n",
        "\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",
        "test_datagen = ImageDataGenerator( rescale = 1.0/255. )\n",
        "\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",
        "validation_generator =  test_datagen.flow_from_directory( validation_dir,\n",
        "                                                          batch_size  = 20,\n",
        "                                                          class_mode  = 'binary', \n",
        "                                                          target_size = (150, 150))"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2020-08-29 03:36:07--  https://storage.googleapis.com/mledu-datasets/cats_and_dogs_filtered.zip\n",
            "Resolving storage.googleapis.com (storage.googleapis.com)... 74.125.200.128, 74.125.68.128, 172.253.118.128, ...\n",
            "Connecting to storage.googleapis.com (storage.googleapis.com)|74.125.200.128|: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  56.8MB/s    in 1.2s    \n",
            "\n",
            "2020-08-29 03:36:09 (56.8 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"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ON4zgWl-Q7pO",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 689
        },
        "outputId": "67ec1c6c-8c39-41c9-c3b1-4721a65cc90f"
      },
      "source": [
        "history = model.fit(\n",
        "            train_generator,\n",
        "            validation_data = validation_generator,\n",
        "            steps_per_epoch = 100,\n",
        "            epochs = 20,\n",
        "            validation_steps = 50,\n",
        "            verbose = 2)"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/20\n",
            "100/100 - 19s - loss: 0.3503 - accuracy: 0.8665 - val_loss: 0.1918 - val_accuracy: 0.9230\n",
            "Epoch 2/20\n",
            "100/100 - 19s - loss: 0.2444 - accuracy: 0.9120 - val_loss: 0.0837 - val_accuracy: 0.9650\n",
            "Epoch 3/20\n",
            "100/100 - 19s - loss: 0.1999 - accuracy: 0.9295 - val_loss: 0.1180 - val_accuracy: 0.9570\n",
            "Epoch 4/20\n",
            "100/100 - 19s - loss: 0.1859 - accuracy: 0.9355 - val_loss: 0.0911 - val_accuracy: 0.9660\n",
            "Epoch 5/20\n",
            "100/100 - 19s - loss: 0.1804 - accuracy: 0.9330 - val_loss: 0.1299 - val_accuracy: 0.9580\n",
            "Epoch 6/20\n",
            "100/100 - 19s - loss: 0.1835 - accuracy: 0.9325 - val_loss: 0.0927 - val_accuracy: 0.9700\n",
            "Epoch 7/20\n",
            "100/100 - 19s - loss: 0.1615 - accuracy: 0.9400 - val_loss: 0.0921 - val_accuracy: 0.9700\n",
            "Epoch 8/20\n",
            "100/100 - 19s - loss: 0.1838 - accuracy: 0.9400 - val_loss: 0.1053 - val_accuracy: 0.9680\n",
            "Epoch 9/20\n",
            "100/100 - 19s - loss: 0.1634 - accuracy: 0.9445 - val_loss: 0.1238 - val_accuracy: 0.9630\n",
            "Epoch 10/20\n",
            "100/100 - 19s - loss: 0.1527 - accuracy: 0.9515 - val_loss: 0.1202 - val_accuracy: 0.9630\n",
            "Epoch 11/20\n",
            "100/100 - 19s - loss: 0.1583 - accuracy: 0.9470 - val_loss: 0.1253 - val_accuracy: 0.9650\n",
            "Epoch 12/20\n",
            "100/100 - 18s - loss: 0.1463 - accuracy: 0.9510 - val_loss: 0.1220 - val_accuracy: 0.9670\n",
            "Epoch 13/20\n",
            "100/100 - 18s - loss: 0.1226 - accuracy: 0.9595 - val_loss: 0.2133 - val_accuracy: 0.9510\n",
            "Epoch 14/20\n",
            "100/100 - 18s - loss: 0.1285 - accuracy: 0.9505 - val_loss: 0.2910 - val_accuracy: 0.9300\n",
            "Epoch 15/20\n",
            "100/100 - 19s - loss: 0.1476 - accuracy: 0.9485 - val_loss: 0.1018 - val_accuracy: 0.9740\n",
            "Epoch 16/20\n",
            "100/100 - 19s - loss: 0.1430 - accuracy: 0.9550 - val_loss: 0.1300 - val_accuracy: 0.9640\n",
            "Epoch 17/20\n",
            "100/100 - 19s - loss: 0.1321 - accuracy: 0.9505 - val_loss: 0.1115 - val_accuracy: 0.9710\n",
            "Epoch 18/20\n",
            "100/100 - 18s - loss: 0.1318 - accuracy: 0.9570 - val_loss: 0.1091 - val_accuracy: 0.9700\n",
            "Epoch 19/20\n",
            "100/100 - 18s - loss: 0.1602 - accuracy: 0.9490 - val_loss: 0.1210 - val_accuracy: 0.9680\n",
            "Epoch 20/20\n",
            "100/100 - 19s - loss: 0.1227 - accuracy: 0.9600 - val_loss: 0.1197 - val_accuracy: 0.9680\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vSNh4Wa7Q-TE",
        "colab_type": "code",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 298
        },
        "outputId": "a58aeebb-f2bc-4ed1-a0d0-986584cb6edf"
      },
      "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()"
      ],
      "execution_count": 6,
      "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": {
            "text/plain": [
              "<Figure size 432x288 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lWPO05KzRAnQ",
        "colab_type": "code",
        "colab": {}
      },
      "source": [
        ""
      ],
      "execution_count": 6,
      "outputs": []
    }
  ]
}