ImageClassificationWithCIFAR_10.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "\n",
        "\n",
        "\n",
        "# **Data pipeline**\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "QCFvszs5fcBt"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "***Downloading and extracting the CIfAR 10 dataset***"
      ],
      "metadata": {
        "id": "Gdt_yVvHgHDM"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8si7-uh8CmMR",
        "outputId": "8c2a970f-6e30-484f-d3c7-44e08edb45b5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2022-11-10 04:28:08--  https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz\n",
            "Resolving www.cs.toronto.edu (www.cs.toronto.edu)... 128.100.3.30\n",
            "Connecting to www.cs.toronto.edu (www.cs.toronto.edu)|128.100.3.30|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 170498071 (163M) [application/x-gzip]\n",
            "Saving to: ‘cifar-10-python.tar.gz’\n",
            "\n",
            "cifar-10-python.tar 100%[===================>] 162.60M  13.8MB/s    in 13s     \n",
            "\n",
            "2022-11-10 04:28:23 (12.1 MB/s) - ‘cifar-10-python.tar.gz’ saved [170498071/170498071]\n",
            "\n"
          ]
        }
      ],
      "source": [
        "!wget https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz\n",
        "!tar -xf /content/cifar-10-python.tar.gz "
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "***Importing the required libraries***"
      ],
      "metadata": {
        "id": "212uud_Vh7XF"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "BSMGZP4TCmMW"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import pickle\n",
        "import numpy as np\n",
        "from PIL import Image\n",
        "import tensorflow as tf\n",
        "import random\n",
        "import time\n",
        "import scipy"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Unpickle the data**"
      ],
      "metadata": {
        "id": "0n4TDPcAiDOJ"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "O5kM1gHTCmMX"
      },
      "outputs": [],
      "source": [
        "def unpickle(file):\n",
        "    import pickle\n",
        "    with open(file, 'rb') as fo:\n",
        "        dict = pickle.load(fo, encoding='bytes')\n",
        "    return dict"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PfeKUIINCmMZ"
      },
      "outputs": [],
      "source": [
        "# The path to data \n",
        "_data_root = \"/content/Data/\""
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Reading the dataset "
      ],
      "metadata": {
        "id": "DdxYdsupiKKQ"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "JAAEOm5GCmMb"
      },
      "outputs": [],
      "source": [
        "def load_data(data_dir):\n",
        "   #load training data \n",
        "    for i in range(1, 6): \n",
        "        filename = data_dir+\"data_batch_\"+str(i) \n",
        "        dictionary = unpickle(filename) \n",
        "        x_data = dictionary[b'data']  \n",
        "        y_data = np.array(dictionary[b\"labels\"])\n",
        "        if i==1:   \n",
        "            x_train = x_data  \n",
        "            y_train= y_data  \n",
        "        else:  \n",
        "            x_train = np.concatenate((x_train, x_data), axis = 0)   \n",
        "            y_train = np.concatenate((y_train, y_data), axis = 0) \n",
        "    #load testing data \n",
        "    filename = data_dir+\"test_batch\" \n",
        "    dictionary = unpickle(filename)\n",
        "    data = dictionary[b\"data\"]\n",
        "    x_test = data \n",
        "    y_test = np.array(dictionary[b\"labels\"]) \n",
        "    return x_train, y_train, x_test, y_test"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Splitting the data to training and validation dataset**"
      ],
      "metadata": {
        "id": "s070vP-siY_P"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "scrolled": true,
        "tags": [],
        "id": "gFQnkY73CmMd"
      },
      "outputs": [],
      "source": [
        "def train_valid_split(x_train, y_train, split_index=48000):\n",
        "    x_train_new = x_train[:split_index] \n",
        "    y_train_new = y_train[:split_index] \n",
        "    x_valid = x_train[split_index:] \n",
        "    y_valid = y_train[split_index:]\n",
        "    return x_train_new, y_train_new, x_valid, y_valid"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**parse our records in the data**:\n",
        "\n",
        "\n",
        "\n",
        "\n",
        "Converting the image format from (n,depth * height * width) to first (n, depth ,height,width) and then to (n, width, height, depth) "
      ],
      "metadata": {
        "id": "vljOJsKCi3yC"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "zhS4c5fACmMe"
      },
      "outputs": [],
      "source": [
        "def parse_record(records):\n",
        "    n = len(records)\n",
        "    depth_major = records.reshape((n,3, 32, 32))\n",
        "    image = np.transpose(depth_major, [0,2, 3, 1])\n",
        "    return image"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "IJJSCrheCmMh"
      },
      "outputs": [],
      "source": [
        "X_train, Y_train, X_test, Y_test = load_data(_data_root)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xQb0UEb8CmMi"
      },
      "outputs": [],
      "source": [
        "X_train_new, Y_train_new, X_valid, Y_valid = train_valid_split(X_train, Y_train)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "scrolled": true,
        "tags": [],
        "id": "amXk7Xx4CmMi"
      },
      "outputs": [],
      "source": [
        "images = parse_record(X_train_new)\n",
        "X_valid = parse_record(X_valid)\n",
        "X_test = parse_record(X_test)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Scale Data**"
      ],
      "metadata": {
        "id": "OSZOvma5jZA4"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "images = images/255\n",
        "X_valid = X_valid/255\n",
        "X_test = X_test/255"
      ],
      "metadata": {
        "id": "05y-mAAXhDnQ"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Converting labels from integer representation to One hot encoding**\n",
        "\n",
        "we will use \"CategoricalCrossentropy()\" loss function which requires its input to be one hot encoded"
      ],
      "metadata": {
        "id": "7dbrSNZdj19U"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from keras.utils import np_utils\n",
        "y_train = np_utils.to_categorical(Y_train_new, 10)\n",
        "y_val = np_utils.to_categorical(Y_valid, 10)\n",
        "Y_test = np_utils.to_categorical(Y_test, 10)"
      ],
      "metadata": {
        "id": "EvJzSiuZg5T1"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "tags": [],
        "id": "X6HbWjtXCmMo"
      },
      "source": [
        "**Showing some data samples**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "EN4T7BImCmMp"
      },
      "outputs": [],
      "source": [
        "from matplotlib import pyplot as plt"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "tags": [],
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 101
        },
        "id": "-WeiwtV2CmMq",
        "outputId": "cd15278f-602d-42b3-d18a-ebdec279a51c"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 2304x2304 with 15 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "fig, ax = plt.subplots(ncols=15, figsize=(32,32))\n",
        "for i in range(15):\n",
        "    ax[i].imshow(images[i])\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "l8Nvvsi1CmMt"
      },
      "source": [
        "#Build Deep Learning Model"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "DVEh5hI1CmMv"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.models import Sequential\n",
        "from tensorflow.keras.layers import Conv2D, MaxPooling2D, Dense, Flatten, Dropout, Activation\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Using data augmentation to prevent the model from Overfitting**"
      ],
      "metadata": {
        "id": "YYK0xcvxlRsA"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "qz8YxlV4CmMw"
      },
      "outputs": [],
      "source": [
        "datagen = ImageDataGenerator(\n",
        "    rotation_range=20,\n",
        "    width_shift_range=0.2,\n",
        "    height_shift_range=0.2,\n",
        "    horizontal_flip=True,\n",
        "    validation_split=0.2)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "**Bulding the model**"
      ],
      "metadata": {
        "id": "NZMH9ddTlq8k"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "0OwhYuI2CmMw"
      },
      "outputs": [],
      "source": [
        "model = Sequential()\n",
        "model.add(Conv2D(32, (3,3), 1,padding=\"same\", activation='relu', input_shape=(32,32,3)))\n",
        "model.add(Conv2D(32, (3,3), 1,padding=\"same\", activation='relu', input_shape=(32,32,3)))\n",
        "model.add(Conv2D(64, (3,3), 1,padding=\"same\", activation='relu'))\n",
        "model.add(Conv2D(64, (3,3), 1,padding=\"same\", activation='relu'))\n",
        "model.add(MaxPooling2D())\n",
        "model.add(Conv2D(128, (3,3), 1, activation='relu'))\n",
        "model.add(MaxPooling2D())\n",
        "model.add(Flatten())\n",
        "model.add(Dense(516, activation='relu'))\n",
        "model.add(Dense(64, activation='relu'))\n",
        "model.add(Dense(10))\n",
        "model.add(Activation('softmax'))\n",
        "model.compile(\"adam\", loss=tf.losses.CategoricalCrossentropy(), metrics=['accuracy'])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "tags": [],
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hVnt1-jPCmMx",
        "outputId": "97b54452-2f17-4ebb-bca5-3504454a6534"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential_2\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " conv2d_10 (Conv2D)          (None, 32, 32, 32)        896       \n",
            "                                                                 \n",
            " conv2d_11 (Conv2D)          (None, 32, 32, 32)        9248      \n",
            "                                                                 \n",
            " conv2d_12 (Conv2D)          (None, 32, 32, 64)        18496     \n",
            "                                                                 \n",
            " conv2d_13 (Conv2D)          (None, 32, 32, 64)        36928     \n",
            "                                                                 \n",
            " max_pooling2d_4 (MaxPooling  (None, 16, 16, 64)       0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " conv2d_14 (Conv2D)          (None, 14, 14, 128)       73856     \n",
            "                                                                 \n",
            " max_pooling2d_5 (MaxPooling  (None, 7, 7, 128)        0         \n",
            " 2D)                                                             \n",
            "                                                                 \n",
            " flatten_2 (Flatten)         (None, 6272)              0         \n",
            "                                                                 \n",
            " dense_6 (Dense)             (None, 516)               3236868   \n",
            "                                                                 \n",
            " dense_7 (Dense)             (None, 64)                33088     \n",
            "                                                                 \n",
            " dense_8 (Dense)             (None, 10)                650       \n",
            "                                                                 \n",
            " activation_2 (Activation)   (None, 10)                0         \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 3,410,030\n",
            "Trainable params: 3,410,030\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model.summary()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MG5AfQHHCmMx"
      },
      "source": [
        "# Train"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "4RH2VZryCmMx"
      },
      "outputs": [],
      "source": [
        "logdir='/content/logs'"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "hEOxN5I_CmMy"
      },
      "outputs": [],
      "source": [
        "tensorboard_callback = tf.keras.callbacks.TensorBoard(log_dir=logdir)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "tags": [],
        "id": "Q0vLrjnoCmMy"
      },
      "outputs": [],
      "source": [
        "hist = model.fit(datagen.flow(images,y_train), epochs=50, validation_data=datagen.flow(X_valid, y_val,\n",
        "         batch_size=8),steps_per_epoch=len(images) / 32, \n",
        "                 callbacks=[tensorboard_callback])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PQvzv_CMCmMy"
      },
      "source": [
        "# Plot Performance"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 294
        },
        "id": "CqGQVBrqCmMz",
        "outputId": "3db48b90-45b1-4e16-9fe8-aa068086aafd"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "fig = plt.figure()\n",
        "plt.plot(hist.history['loss'], color='teal', label='loss')\n",
        "plt.plot(hist.history['val_loss'], color='orange', label='val_loss')\n",
        "fig.suptitle('Loss', fontsize=20)\n",
        "plt.legend(loc=\"upper left\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 294
        },
        "id": "Au78v6O7CmMz",
        "outputId": "c3ee4999-0d65-43f8-cd7f-e294fe8c18ff"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "fig = plt.figure()\n",
        "plt.plot(hist.history['accuracy'], color='teal', label='accuracy')\n",
        "plt.plot(hist.history['val_accuracy'], color='orange', label='val_accuracy')\n",
        "fig.suptitle('Accuracy', fontsize=20)\n",
        "plt.legend(loc=\"upper left\")\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Save the Model\n"
      ],
      "metadata": {
        "id": "Hr2Uq_dPmRmB"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model.save(os.path.join('models','imageclassifier.h5'))"
      ],
      "metadata": {
        "id": "faH_Q74rmd25"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZlunhVndCmMz"
      },
      "source": [
        "# Evaluate"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "IdStei_HCmM0"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.metrics import BinaryAccuracy"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "a8gMPDAUCmM0"
      },
      "outputs": [],
      "source": [
        "acc = BinaryAccuracy()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "U3-LH9t9CmM1"
      },
      "outputs": [],
      "source": [
        "test_dataset = tf.data.Dataset.from_tensor_slices((X_test, Y_test))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0eatPh0VCmM2",
        "outputId": "9b915d16-bae6-4025-b90e-632d7a783abd"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 37s 119ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Variable 'UnreadVariable' shape=() dtype=float32, numpy=109980.0>"
            ]
          },
          "metadata": {},
          "execution_count": 70
        }
      ],
      "source": [
        "yhat = model.predict(X_test)\n",
        "pre.update_state(Y_test, yhat)\n",
        "re.update_state(Y_test, yhat)\n",
        "acc.update_state(Y_test, yhat)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IWwrFcMjCmM2",
        "outputId": "7d62f6ba-7939-439a-ea1b-9066514f4ca8"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            " tf.Tensor(0.96031094, shape=(), dtype=float32)\n"
          ]
        }
      ],
      "source": [
        "print(acc.result())"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cr5vdH30CmM2"
      },
      "source": [
        "# Test"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "GQrEPqIhCmM2"
      },
      "outputs": [],
      "source": [
        "img = X_test[600]"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.models import load_model"
      ],
      "metadata": {
        "id": "T5IrldAcm12W"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "new_model = load_model('/content/models/imageclassifier.h5')"
      ],
      "metadata": {
        "id": "H3a8aqpfm94z"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "new_model.predict(np.expand_dims(img, 0))"
      ],
      "metadata": {
        "id": "ZRiTJJ0wnD6c"
      },
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    },
    "colab": {
      "provenance": [],
      "collapsed_sections": []
    },
    "gpuClass": "standard"
  },
  "nbformat": 4,
  "nbformat_minor": 0
}