nnproject.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "nnproject.ipynb",
      "provenance": [],
      "authorship_tag": "ABX9TyN7U9XqWxts2ocNTEPBREY2"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "code",
      "metadata": {
        "id": "BRA-7BrfzQFM"
      },
      "source": [
        "import numpy as np\n",
        "import pandas as pd\n",
        "import matplotlib.pyplot as plt\n",
        "%matplotlib inline\n",
        "from keras.datasets import mnist\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Dense, Dropout, Activation, Flatten\n",
        "from keras.optimizers import Adam\n",
        "from keras.layers.normalization import BatchNormalization\n",
        "from keras.utils import np_utils\n",
        "from keras.layers import Conv2D, MaxPooling2D, ZeroPadding2D, GlobalAveragePooling2D\n",
        "from keras.layers.advanced_activations import LeakyReLU \n",
        "from keras.preprocessing.image import ImageDataGenerator\n",
        "\n",
        "np.random.seed(25)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "N6CLCfZizhoi",
        "outputId": "a6e421b6-c2cf-4c75-82fa-6f7e752abf60",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 118
        }
      },
      "source": [
        "(X_train, y_train), (X_test, y_test) = mnist.load_data()\n",
        "print(\"X_train original shape\", X_train.shape)\n",
        "print(\"y_train original shape\", y_train.shape)\n",
        "print(\"X_test original shape\", X_test.shape)\n",
        "print(\"y_test original shape\", y_test.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
            "11493376/11490434 [==============================] - 0s 0us/step\n",
            "X_train original shape (60000, 28, 28)\n",
            "y_train original shape (60000,)\n",
            "X_test original shape (10000, 28, 28)\n",
            "y_test original shape (10000,)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dg4zbfMQzuq9",
        "outputId": "0786452e-7770-4a91-fd57-d389250a6562",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 302
        }
      },
      "source": [
        "# summarize loaded dataset\n",
        "print('Train: X=%s, y=%s' % (X_train.shape, y_train.shape))\n",
        "print('Test: X=%s, y=%s' % (X_test.shape, y_test.shape))\n",
        "\n",
        "#first few data of mnist\n",
        "for i in range(9):\n",
        "\t# define subplot\n",
        "\tplt.subplot(330 + 1 + i)\n",
        "\t# plot raw pixel data\n",
        "\tplt.imshow(X_train[i], cmap=plt.get_cmap('gray'))\n",
        "# show the figure\n",
        "plt.show()\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Train: X=(60000, 28, 28), y=(60000,)\n",
            "Test: X=(10000, 28, 28), y=(10000,)\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 9 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iSUa5-EBz54q",
        "outputId": "9d12cb5e-454d-4109-924b-048155352abb",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "X_train = X_train.reshape(X_train.shape[0], 28, 28, 1)\n",
        "X_test = X_test.reshape(X_test.shape[0], 28, 28, 1)\n",
        "\n",
        "X_train = X_train.astype('float32')\n",
        "X_test = X_test.astype('float32')\n",
        "\n",
        "X_train/=255\n",
        "X_test/=255\n",
        "\n",
        "X_train.shape"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000, 28, 28, 1)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 9
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Qlc4HMgtz90Q",
        "outputId": "6b64e12d-daa3-4c9d-f909-1d7db521270f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 34
        }
      },
      "source": [
        "number_of_classes = 10\n",
        "\n",
        "Y_train = np_utils.to_categorical(y_train, number_of_classes)\n",
        "Y_test = np_utils.to_categorical(y_test, number_of_classes)\n",
        "\n",
        "y_train[0], Y_train[0]"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(5, array([0., 0., 0., 0., 0., 1., 0., 0., 0., 0.], dtype=float32))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "OOF6ycYo0B7G"
      },
      "source": [
        "# Three steps to Convolution\n",
        "# 1. Convolution\n",
        "# 2. Activation\n",
        "# 3. Polling\n",
        "# Repeat Steps 1,2,3 for adding more hidden layers\n",
        "\n",
        "# 4. After that make a fully connected network\n",
        "# This fully connected network gives ability to the CNN\n",
        "# to classify the samples\n",
        "\n",
        "model = Sequential()\n",
        "\n",
        "model.add(Conv2D(32, (3, 3), input_shape=(28,28,1)))\n",
        "model.add(Activation('relu'))\n",
        "BatchNormalization(axis=-1)\n",
        "model.add(Conv2D(32, (3, 3)))\n",
        "model.add(Activation('relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2,2)))\n",
        "\n",
        "BatchNormalization(axis=-1)\n",
        "model.add(Conv2D(64,(3, 3)))\n",
        "model.add(Activation('relu'))\n",
        "BatchNormalization(axis=-1)\n",
        "model.add(Conv2D(64, (3, 3)))\n",
        "model.add(Activation('relu'))\n",
        "model.add(MaxPooling2D(pool_size=(2,2)))\n",
        "\n",
        "model.add(Flatten())\n",
        "# Fully connected layer\n",
        "\n",
        "BatchNormalization()\n",
        "model.add(Dense(512))\n",
        "model.add(Activation('relu'))\n",
        "BatchNormalization()\n",
        "model.add(Dropout(0.2))\n",
        "model.add(Dense(10))\n",
        "\n",
        "# model.add(Convolution2D(10,3,3, border_mode='same'))\n",
        "# model.add(GlobalAveragePooling2D())\n",
        "model.add(Activation('softmax'))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7o41Fnwx0GLT",
        "outputId": "e2a59b0c-6bb4-42bd-a84f-14e922c0cf89",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 689
        }
      },
      "source": [
        "model.summary()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"sequential\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "conv2d (Conv2D)              (None, 26, 26, 32)        320       \n",
            "_________________________________________________________________\n",
            "activation (Activation)      (None, 26, 26, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_1 (Conv2D)            (None, 24, 24, 32)        9248      \n",
            "_________________________________________________________________\n",
            "activation_1 (Activation)    (None, 24, 24, 32)        0         \n",
            "_________________________________________________________________\n",
            "max_pooling2d (MaxPooling2D) (None, 12, 12, 32)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_2 (Conv2D)            (None, 10, 10, 64)        18496     \n",
            "_________________________________________________________________\n",
            "activation_2 (Activation)    (None, 10, 10, 64)        0         \n",
            "_________________________________________________________________\n",
            "conv2d_3 (Conv2D)            (None, 8, 8, 64)          36928     \n",
            "_________________________________________________________________\n",
            "activation_3 (Activation)    (None, 8, 8, 64)          0         \n",
            "_________________________________________________________________\n",
            "max_pooling2d_1 (MaxPooling2 (None, 4, 4, 64)          0         \n",
            "_________________________________________________________________\n",
            "flatten (Flatten)            (None, 1024)              0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 512)               524800    \n",
            "_________________________________________________________________\n",
            "activation_4 (Activation)    (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 512)               0         \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 10)                5130      \n",
            "_________________________________________________________________\n",
            "activation_5 (Activation)    (None, 10)                0         \n",
            "=================================================================\n",
            "Total params: 594,922\n",
            "Trainable params: 594,922\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A3lzoiT-0KuJ"
      },
      "source": [
        "model.compile(loss='categorical_crossentropy', optimizer=Adam(), metrics=['accuracy'])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "j9_hw3Kc0Qc6"
      },
      "source": [
        "gen = ImageDataGenerator(rotation_range=8, width_shift_range=0.08, shear_range=0.3,\n",
        "                         height_shift_range=0.08, zoom_range=0.08)\n",
        "\n",
        "test_gen = ImageDataGenerator()"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "GRzEi8180UCd"
      },
      "source": [
        "train_generator = gen.flow(X_train, Y_train, batch_size=64)\n",
        "test_generator = test_gen.flow(X_test, Y_test, batch_size=64)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lDy0fQP30XF6",
        "outputId": "a122ea56-5d68-4f32-a85b-58c87dda1a1f",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 185
        }
      },
      "source": [
        "# model.fit(X_train, Y_train, batch_size=128, nb_epoch=1, validation_data=(X_test, Y_test))\n",
        "\n",
        "history = model.fit(train_generator, steps_per_epoch=60000//64, epochs=5, \n",
        "                    validation_data=test_generator, validation_steps=10000//64)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/5\n",
            "937/937 [==============================] - 27s 29ms/step - loss: 0.2022 - accuracy: 0.9361 - val_loss: 0.0342 - val_accuracy: 0.9878\n",
            "Epoch 2/5\n",
            "937/937 [==============================] - 27s 29ms/step - loss: 0.0658 - accuracy: 0.9792 - val_loss: 0.0323 - val_accuracy: 0.9893\n",
            "Epoch 3/5\n",
            "937/937 [==============================] - 27s 29ms/step - loss: 0.0488 - accuracy: 0.9850 - val_loss: 0.0248 - val_accuracy: 0.9917\n",
            "Epoch 4/5\n",
            "937/937 [==============================] - 27s 29ms/step - loss: 0.0394 - accuracy: 0.9878 - val_loss: 0.0199 - val_accuracy: 0.9934\n",
            "Epoch 5/5\n",
            "937/937 [==============================] - 27s 29ms/step - loss: 0.0371 - accuracy: 0.9886 - val_loss: 0.0240 - val_accuracy: 0.9920\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "mu9-Ps2a0jOJ",
        "outputId": "36227568-6945-4f5f-8455-535f1fde8c71",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 67
        }
      },
      "source": [
        "score = model.evaluate(X_test, Y_test)\n",
        "print()\n",
        "print('Test accuracy: ', score[1])"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "313/313 [==============================] - 1s 3ms/step - loss: 0.0172 - accuracy: 0.9951\n",
            "\n",
            "Test accuracy:  0.9951000213623047\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "p8DPWeGs1rEB"
      },
      "source": [
        "def plot_curve(epochs, hist, list_of_metrics):\n",
        "  plt.figure()\n",
        "  plt.xlabel(\"Epoch\")\n",
        "  plt.ylabel(\"Value\")\n",
        "\n",
        "  for m in list_of_metrics:\n",
        "    x = hist[m]\n",
        "    plt.plot(epochs[1:],x[1:], label = m)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1nEMENkI11vk",
        "outputId": "09be16f4-1820-4857-912d-7934711f82d8",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        }
      },
      "source": [
        "  epochs= history.epoch\n",
        "  hist = pd.DataFrame(history.history)\n",
        "  list_of_metrics_to_plot = ['accuracy']\n",
        "  plot_curve(epochs, hist, list_of_metrics_to_plot)"
      ],
      "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": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    }
  ]
}