Transfer_Learning_VGG19.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Transfer_Learning_VGG19.ipynb",
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GJ22lZAexGvv",
        "outputId": "5093121e-d00e-4c7f-853b-7ebf82a7a090"
      },
      "source": [
        "!pwd"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pw9dl10PxJm0",
        "outputId": "c8d03f9f-4a39-4efe-a9af-495218430ef8"
      },
      "source": [
        "#changing the working directory and entering to dataset directory for my google drive\n",
        "%cd drive/MyDrive/kaggle/image_classification/"
      ],
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content/drive/MyDrive/kaggle/image_classification\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V_Z0jHKmxJqK",
        "outputId": "b9859547-fdb9-4b66-ae2b-bb0bc246f358"
      },
      "source": [
        "%cd flower_classification/"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "/content/drive/MyDrive/kaggle/image_classification/flower_classification\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OmPKzQo-xJt6",
        "outputId": "42e5c6ca-3773-48dc-fb51-5312b95d2a18"
      },
      "source": [
        "!ls"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "AccVal_acc.png\tLossVal_loss.png  test\ttrain\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ahgaq74exJwg"
      },
      "source": [
        "from keras.layers import Input, Lambda, Dense, Flatten\n",
        "from keras.models import Model\n",
        "from keras.applications.vgg19 import VGG19\n",
        "from keras.applications.vgg19 import preprocess_input\n",
        "from keras.preprocessing import image\n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "from keras.models import Sequential\n",
        "import numpy as np\n",
        "from glob import glob\n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lTzdVuBcxJ1L"
      },
      "source": [
        "# re-size all the images to this\n",
        "IMAGE_SIZE = [224, 224]"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "URME5P_kxJ4D"
      },
      "source": [
        "train_path = 'train'\n",
        "valid_path = 'test'"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-qga9FdxxJ8t",
        "outputId": "87ceb540-71b2-4c71-e68e-366bde3c83a5"
      },
      "source": [
        "# add preprocessing layer to the front of VGG19\n",
        "vgg = VGG19(input_shape=IMAGE_SIZE + [3], weights='imagenet', include_top=False)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/vgg19/vgg19_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "80142336/80134624 [==============================] - 1s 0us/step\n",
            "80150528/80134624 [==============================] - 1s 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "c0t12FApxKCd"
      },
      "source": [
        "# don't train existing weights\n",
        "for layer in vgg.layers:\n",
        "  layer.trainable = False"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Z_narI0oxKIz"
      },
      "source": [
        "# useful for getting number of classes\n",
        "folders = glob('train/*')"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dar6MWR_xKMz"
      },
      "source": [
        "# our layers - you can add more if you want\n",
        "x = Flatten()(vgg.output)\n",
        "# x = Dense(1000, activation='relu')(x)\n",
        "prediction = Dense(len(folders), activation='softmax')(x)"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Bv2UnJAWxkB9"
      },
      "source": [
        "# create a model object\n",
        "model = Model(inputs=vgg.input, outputs=prediction)"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "r7xUyDVGxkET",
        "outputId": "51c3ee2f-b376-4768-c7b4-91594bf2b463"
      },
      "source": [
        "# view the structure of the model\n",
        "model.summary()"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_1 (InputLayer)         [(None, 224, 224, 3)]     0         \n",
            "_________________________________________________________________\n",
            "block1_conv1 (Conv2D)        (None, 224, 224, 64)      1792      \n",
            "_________________________________________________________________\n",
            "block1_conv2 (Conv2D)        (None, 224, 224, 64)      36928     \n",
            "_________________________________________________________________\n",
            "block1_pool (MaxPooling2D)   (None, 112, 112, 64)      0         \n",
            "_________________________________________________________________\n",
            "block2_conv1 (Conv2D)        (None, 112, 112, 128)     73856     \n",
            "_________________________________________________________________\n",
            "block2_conv2 (Conv2D)        (None, 112, 112, 128)     147584    \n",
            "_________________________________________________________________\n",
            "block2_pool (MaxPooling2D)   (None, 56, 56, 128)       0         \n",
            "_________________________________________________________________\n",
            "block3_conv1 (Conv2D)        (None, 56, 56, 256)       295168    \n",
            "_________________________________________________________________\n",
            "block3_conv2 (Conv2D)        (None, 56, 56, 256)       590080    \n",
            "_________________________________________________________________\n",
            "block3_conv3 (Conv2D)        (None, 56, 56, 256)       590080    \n",
            "_________________________________________________________________\n",
            "block3_conv4 (Conv2D)        (None, 56, 56, 256)       590080    \n",
            "_________________________________________________________________\n",
            "block3_pool (MaxPooling2D)   (None, 28, 28, 256)       0         \n",
            "_________________________________________________________________\n",
            "block4_conv1 (Conv2D)        (None, 28, 28, 512)       1180160   \n",
            "_________________________________________________________________\n",
            "block4_conv2 (Conv2D)        (None, 28, 28, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "block4_conv3 (Conv2D)        (None, 28, 28, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "block4_conv4 (Conv2D)        (None, 28, 28, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "block4_pool (MaxPooling2D)   (None, 14, 14, 512)       0         \n",
            "_________________________________________________________________\n",
            "block5_conv1 (Conv2D)        (None, 14, 14, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "block5_conv2 (Conv2D)        (None, 14, 14, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "block5_conv3 (Conv2D)        (None, 14, 14, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "block5_conv4 (Conv2D)        (None, 14, 14, 512)       2359808   \n",
            "_________________________________________________________________\n",
            "block5_pool (MaxPooling2D)   (None, 7, 7, 512)         0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 25088)             0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 5)                 125445    \n",
            "=================================================================\n",
            "Total params: 20,149,829\n",
            "Trainable params: 125,445\n",
            "Non-trainable params: 20,024,384\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-hynPpBMxkGo"
      },
      "source": [
        "# tell the model what cost and optimization method to use\n",
        "model.compile(\n",
        "  loss='categorical_crossentropy',\n",
        "  optimizer='adam',\n",
        "  metrics=['accuracy']\n",
        ")"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4h3MQqk6xkI9"
      },
      "source": [
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "\n",
        "train_datagen = ImageDataGenerator(rescale = 1./255,\n",
        "                                   shear_range = 0.2,\n",
        "                                   zoom_range = 0.2,\n",
        "                                   horizontal_flip = True)\n",
        "\n",
        "test_datagen = ImageDataGenerator(rescale = 1./255)"
      ],
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ENIz3T5CxkLK",
        "outputId": "d42b1055-8d07-42b3-aae4-15d5cef54631"
      },
      "source": [
        "training_set = train_datagen.flow_from_directory('train',\n",
        "                                                 target_size = (224, 224),\n",
        "                                                 batch_size = 64,\n",
        "                                                 class_mode = 'categorical')"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 3027 images belonging to 5 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TU16uQKOxkQg",
        "outputId": "c01b304f-a455-4e9f-8c53-7d566f416fbd"
      },
      "source": [
        "test_set = test_datagen.flow_from_directory('test',\n",
        "                                            target_size = (224, 224),\n",
        "                                            batch_size = 64,\n",
        "                                            class_mode = 'categorical')"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Found 1296 images belonging to 5 classes.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IDXQfpgBxw0M",
        "outputId": "5537aa3e-0a5c-40fa-e0f3-d228a07a1c47"
      },
      "source": [
        "# fit the model\n",
        "r = model.fit_generator(\n",
        "  training_set,\n",
        "  validation_data=test_set,\n",
        "  epochs=4,\n",
        "  steps_per_epoch=len(training_set),\n",
        "  validation_steps=len(test_set)\n",
        ")"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/engine/training.py:1972: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
            "  warnings.warn('`Model.fit_generator` is deprecated and '\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/4\n",
            "48/48 [==============================] - 2265s 47s/step - loss: 1.2452 - accuracy: 0.5808 - val_loss: 0.8576 - val_accuracy: 0.7029\n",
            "Epoch 2/4\n",
            "48/48 [==============================] - 61s 1s/step - loss: 0.5993 - accuracy: 0.7958 - val_loss: 0.6530 - val_accuracy: 0.7739\n",
            "Epoch 3/4\n",
            "48/48 [==============================] - 61s 1s/step - loss: 0.4673 - accuracy: 0.8414 - val_loss: 0.7209 - val_accuracy: 0.7515\n",
            "Epoch 4/4\n",
            "48/48 [==============================] - 60s 1s/step - loss: 0.4039 - accuracy: 0.8702 - val_loss: 0.6486 - val_accuracy: 0.7755\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "pT0fEecXxw2j",
        "outputId": "2fda7551-c2ac-4481-b2d2-7d11a6042bb5"
      },
      "source": [
        "# loss\n",
        "plt.plot(r.history['loss'], label='train loss')\n",
        "plt.plot(r.history['val_loss'], label='val loss')\n",
        "plt.legend()\n",
        "plt.show()\n",
        "plt.savefig('LossVal_loss')"
      ],
      "execution_count": 19,
      "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",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 282
        },
        "id": "9qolv9orxw5K",
        "outputId": "eb2cca30-94d3-4e71-bcfb-88f0e856f6c4"
      },
      "source": [
        "# accuracies\n",
        "plt.plot(r.history['accuracy'], label='train acc')\n",
        "plt.plot(r.history['val_accuracy'], label='val acc')\n",
        "plt.legend()\n",
        "plt.show()\n",
        "plt.savefig('AccVal_acc')"
      ],
      "execution_count": 20,
      "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",
      "metadata": {
        "id": "4JhHAfkyxw-q"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}