Cifar10Classification.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Cifar10Classification.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "DL_VHJZ8W-Rd"
      },
      "source": [
        "from keras import layers \n",
        "from keras import models \n",
        "from keras import optimizers \n",
        "from keras.preprocessing.image import ImageDataGenerator \n",
        "from keras.applications.vgg16 import VGG16\n",
        "from keras.preprocessing import image\n",
        "from keras.applications.imagenet_utils import preprocess_input\n",
        "from keras.layers import Input, Flatten, Dense, Conv2D, ActivityRegularization\n",
        "from keras.models import Model\n",
        "import numpy as np\n",
        "from keras import backend as k\n",
        "import tensorflow as tf\n",
        "from keras.applications import resnet50\n",
        "import os \n",
        "import matplotlib.pyplot as plt"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kwN3XQL1SDGM"
      },
      "source": [
        "##Unzip the Cifar dataset "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "29jMEff-jGWB",
        "outputId": "e194cc7a-6148-4586-ab46-8f8508292088",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "!p7zip -d Cifar4_SortedImages.7z "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "/usr/bin/p7zip: cannot read Cifar4_SortedImages.7z\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HDZvZDgRzpbV"
      },
      "source": [
        ""
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yFJhPD9WSLxI"
      },
      "source": [
        "##Model Structure. \n",
        "####----> CNN layers:- \n",
        "These are typically used to reduce the dimensionality of the network. Features of a fully connected layer. Fully connected layers are placed before the classification output of a CNN and are used to flatten the results before classification.\n",
        "\n",
        "####----> Max Pooling layers:-\n",
        "Pooling layers are used to reduce the dimensions of the feature maps. Thus, it reduces the number of parameters to learn and the amount of computation performed in the network. The pooling layer summarises the features present in a region of the feature map generated by a convolution layer.\n",
        "\n",
        "####----> Dense Layers :- \n",
        "Dense layers are used when association can exist among any feature to any other feature in data point.\n",
        "\n",
        "####----> ReLu:-\n",
        "ReLU is important because it does not saturate; the gradient is always high (equal to 1) if the neuron activates. As long as it is not a dead neuron, successive updates are fairly effective. ReLU is also very quick to evaluate.\n",
        "\n",
        "####----> Softmax:- \n",
        "Softmax is very useful because it converts the scores to a normalized probability distribution, which can be displayed to a user or used as input to other systems. For this reason it is usual to append a softmax function as the final layer of the neural network."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5kAu9FoBXUMt"
      },
      "source": [
        "def MakeModel(weights_path=None):\n",
        "    model = models.Sequential() \n",
        "    model.add(layers.Conv2D(64, (3, 3), activation='relu', input_shape=(32, 32, 3))) \n",
        "    model.add(layers.MaxPooling2D((2, 2))) \n",
        "    #-----------------------------------\n",
        "    #-------Build your layers here -----\n",
        "    model.add(layers.Conv2D(128, (3, 3), activation='relu', input_shape=(32, 32, 3))) \n",
        "    model.add(layers.MaxPooling2D((2, 2)))   \n",
        "    model.add(layers.Dense(1024,activation='relu'))\n",
        "    model.add(layers.Dropout(0.5))\n",
        "    model.add(layers.Dense(512,activation='relu'))\n",
        "    model.add(layers.Dense(256,activation='tanh'))\n",
        "    #-----------------------------------\n",
        "    \n",
        "    model.add(layers.Flatten()) \n",
        "    model.add(layers.Dropout(0.5))\n",
        "    model.add(layers.Dense(512, activation='relu')) \n",
        "    model.add(layers.Dense(4, activation='softmax', name='output')) \n",
        "    #model.add(layers.Dense(1000, activation='softmax'))\n",
        "    return model"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Mi9LchtLykvZ"
      },
      "source": [
        "###Calculating cross entropy loss function and we are using Adam optimizer to optimize our network over time.\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "0TTdz9dpXUIi",
        "outputId": "8674f8cb-f287-4675-d426-52180985b3af",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 561
        }
      },
      "source": [
        "from keras.models import Model, load_model\n",
        "from tensorflow.keras import regularizers\n",
        "\n",
        "base_dir = 'Cifar4_SortedImages' \n",
        "train_dir = os.path.join(base_dir, 'train') \n",
        "validation_dir = os.path.join(base_dir, 'validation') \n",
        "test_dir = os.path.join(base_dir, 'test') \n",
        "\n",
        "model = MakeModel()\n",
        "\n",
        "# Displaying a summary of the model \n",
        "model.summary() \n",
        "\n",
        "# Configuring our model for training \n",
        "# For a binary classification problem \n",
        "model.compile(loss='categorical_crossentropy', optimizer=optimizers.Adam(lr=0.0001), metrics=['accuracy']) "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 30, 30, 64)        1792      \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 15, 15, 64)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 13, 13, 128)       73856     \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 6, 6, 128)         0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 6, 6, 1024)        132096    \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 6, 6, 1024)        0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 6, 6, 512)         524800    \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 6, 6, 256)         131328    \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 9216)              0         \n",
            "_________________________________________________________________\n",
            "dropout_1 (Dropout)          (None, 9216)              0         \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 512)               4719104   \n",
            "_________________________________________________________________\n",
            "output (Dense)               (None, 4)                 2052      \n",
            "=================================================================\n",
            "Total params: 5,585,028\n",
            "Trainable params: 5,585,028\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2Xjr2oEqy_53"
      },
      "source": [
        "#Training \n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Tu4t33WBXUF1",
        "outputId": "27e5f60e-5867-451d-ca37-a4a88033011d",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 527
        }
      },
      "source": [
        "\n",
        "# Using ImageDataGenerator to read images from directories \n",
        "# all images will be rescaled by 1./255 \n",
        "#train_datagen = ImageDataGenerator(rescale=1./255)  \n",
        "\n",
        "#preprocess_input = resnet50.preprocess_input\n",
        "train_datagen = ImageDataGenerator(\n",
        "    rescale=1./255,\n",
        "    preprocessing_function=preprocess_input\n",
        ")\n",
        "\n",
        "validation_datagen = ImageDataGenerator(rescale=1./255, preprocessing_function=preprocess_input) \n",
        "test_datagen = ImageDataGenerator(rescale=1./255, preprocessing_function=preprocess_input) \n",
        "\n",
        "train_generator = train_datagen.flow_from_directory( \n",
        "    train_dir, # this is the target directory \n",
        "    target_size=(32, 32),  \n",
        "    batch_size=32,\n",
        "    shuffle=True,\n",
        "    class_mode = 'categorical') \n",
        "    # since we use categorical_crossentropy loss, we need categorical labels \n",
        " \n",
        "validation_generator = validation_datagen.flow_from_directory( \n",
        "    validation_dir, \n",
        "    target_size=(32, 32), \n",
        "    batch_size=32,\n",
        "    shuffle=True,\n",
        "    class_mode = 'categorical')\n",
        "\n",
        "step_size_train=train_generator.n//train_generator.batch_size\n",
        "step_size_val = validation_generator.samples // validation_generator.batch_size\n",
        "\n",
        "from keras.callbacks import EarlyStopping, ModelCheckpoint\n",
        "\n",
        "checkpointer = ModelCheckpoint(\n",
        "\t\tfilepath='weights.h5',\n",
        "\t\tverbose=1, save_best_only=True\t\n",
        "\t)\n",
        "\n",
        "early_stopping = EarlyStopping(monitor='val_loss',patience=10)\n",
        "\n",
        "\n",
        "# Fitting our model using a batch generator \n",
        "# Trains the model for a fixed number of epochs (iterations on a dataset). \n",
        "history = model.fit_generator( \n",
        "   train_generator, \n",
        "   steps_per_epoch=step_size_train, \n",
        "   epochs=10,  \n",
        "   # callbacks=[checkpointer,early_stopping],\n",
        "   validation_data=validation_generator, \n",
        "   validation_steps=step_size_val) \n",
        " \n",
        " \n",
        "# Evaluating the model - test on finalised model \n",
        "# Returns the loss value & metrics values for the model in test mode. \n",
        "test_generator = test_datagen.flow_from_directory( \n",
        "    test_dir, \n",
        "    target_size=(32, 32), \n",
        "    batch_size=32,\n",
        "    shuffle=True,\n",
        "    class_mode = 'categorical') \n",
        " \n",
        "\n",
        "step_size_test = test_generator.samples // test_generator.batch_size\n",
        "# finally evaluate this model on the test data \n",
        "results = model.evaluate_generator( \n",
        "    test_generator, \n",
        "    steps=step_size_test) \n",
        "print('Final test accuracy:', (results[1]*100.0))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Found 15000 images belonging to 4 classes.\n",
            "Found 3000 images belonging to 4 classes.\n",
            "WARNING:tensorflow:From <ipython-input-5-d752010260ca>:51: Model.fit_generator (from tensorflow.python.keras.engine.training) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Please use Model.fit, which supports generators.\n",
            "Epoch 1/10\n",
            "468/468 [==============================] - 8s 17ms/step - loss: 1.1726 - accuracy: 0.4735 - val_loss: 1.0366 - val_accuracy: 0.5514\n",
            "Epoch 2/10\n",
            "468/468 [==============================] - 8s 17ms/step - loss: 0.9833 - accuracy: 0.5829 - val_loss: 0.9240 - val_accuracy: 0.6203\n",
            "Epoch 3/10\n",
            "468/468 [==============================] - 8s 17ms/step - loss: 0.9021 - accuracy: 0.6235 - val_loss: 0.9091 - val_accuracy: 0.6196\n",
            "Epoch 4/10\n",
            "468/468 [==============================] - 8s 18ms/step - loss: 0.8398 - accuracy: 0.6518 - val_loss: 0.8285 - val_accuracy: 0.6620\n",
            "Epoch 5/10\n",
            "468/468 [==============================] - 8s 17ms/step - loss: 0.7851 - accuracy: 0.6764 - val_loss: 0.8225 - val_accuracy: 0.6677\n",
            "Epoch 6/10\n",
            "468/468 [==============================] - 8s 17ms/step - loss: 0.7359 - accuracy: 0.6986 - val_loss: 0.7693 - val_accuracy: 0.6895\n",
            "Epoch 7/10\n",
            "468/468 [==============================] - 8s 17ms/step - loss: 0.6924 - accuracy: 0.7159 - val_loss: 0.7639 - val_accuracy: 0.6899\n",
            "Epoch 8/10\n",
            "468/468 [==============================] - 8s 17ms/step - loss: 0.6514 - accuracy: 0.7396 - val_loss: 0.7380 - val_accuracy: 0.7009\n",
            "Epoch 9/10\n",
            "468/468 [==============================] - 8s 17ms/step - loss: 0.6097 - accuracy: 0.7570 - val_loss: 0.7227 - val_accuracy: 0.7174\n",
            "Epoch 10/10\n",
            "468/468 [==============================] - 8s 17ms/step - loss: 0.5720 - accuracy: 0.7728 - val_loss: 0.7166 - val_accuracy: 0.7144\n",
            "Found 2000 images belonging to 4 classes.\n",
            "WARNING:tensorflow:From <ipython-input-5-d752010260ca>:68: Model.evaluate_generator (from tensorflow.python.keras.engine.training) is deprecated and will be removed in a future version.\n",
            "Instructions for updating:\n",
            "Please use Model.evaluate, which supports generators.\n",
            "Final test accuracy: 69.65726017951965\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_Y6Rm5hezOME"
      },
      "source": [
        "#Plotting both Training versus Validation for Accuracy and Loss"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S8Fp11h1XUDM",
        "outputId": "a5d34296-fd94-4dcc-913e-d23ea43d9169",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 590
        }
      },
      "source": [
        "print(history.history.keys())\n",
        "#  \"Accuracy\"\n",
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.title('model accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'validation'], loc='upper left')\n",
        "plt.show()\n",
        "# \"Loss\"\n",
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history.history['val_loss'])\n",
        "plt.title('model loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['train', 'validation'], loc='upper left')\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "dict_keys(['loss', 'accuracy', 'val_loss', 'val_accuracy'])\n"
          ],
          "name": "stdout"
        },
        {
          "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"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mtTeeRy_XT9y"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}