005/005-cnn-mnist.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Классификация MNIST сверточной сетью\n",
    "\n",
    "https://www.kaggle.com/c/digit-recognizer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "2.2.0\n",
      "True\n"
     ]
    }
   ],
   "source": [
    "import cv2\n",
    "import numpy as np\n",
    "import tensorflow as tf\n",
    "\n",
    "print(tf.__version__)\n",
    "print(tf.executing_eagerly())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "train = np.loadtxt('../002/data/digit/train.csv', delimiter=',', skiprows=1)\n",
    "test = np.loadtxt('../002/data/digit/test.csv', delimiter=',', skiprows=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# сохраняем разметку в отдельную переменную\n",
    "train_label = train[:, 0]\n",
    "\n",
    "# приводим размерность к удобному для обаботки виду\n",
    "# добавляем размерность канала\n",
    "train_img = np.resize(train[:, 1:], (train.shape[0], 28, 28, 1))\n",
    "test_img = np.resize(test, (test.shape[0], 28, 28, 1))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Визуализируем исходные данные"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "%matplotlib inline\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "fig = plt.figure(figsize=(20, 10))\n",
    "for i, img in enumerate(train_img[0:5, :], 1):\n",
    "    subplot = fig.add_subplot(1, 5, i)\n",
    "    plt.imshow(img[:,:,0], cmap='gray');\n",
    "    subplot.set_title('%s' % train_label[i - 1]);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Разбиваем выборку на обучение и валидацию"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import train_test_split\n",
    "y_train, y_val, x_train, x_val = train_test_split(\n",
    "    train_label, train_img, test_size=0.2, random_state=42)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Собираем сверточную сеть для обучения"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = tf.keras.models.Sequential([\n",
    "    tf.keras.layers.Conv2D(filters=6,\n",
    "                           kernel_size=(5, 5),\n",
    "                           padding='same',\n",
    "                           activation='relu',\n",
    "                           input_shape=x_train.shape[1:]),\n",
    "    tf.keras.layers.MaxPool2D(pool_size=(2, 2), padding='valid'),\n",
    "    \n",
    "    tf.keras.layers.Conv2D(filters=16, \n",
    "                           kernel_size=(5, 5),\n",
    "                           padding='same',\n",
    "                           activation='relu'),\n",
    "    tf.keras.layers.MaxPool2D(pool_size=(2, 2), padding='valid'),\n",
    "    \n",
    "    tf.keras.layers.Flatten(),\n",
    "    tf.keras.layers.Dense(32, activation='relu'),\n",
    "    tf.keras.layers.Dense(10, activation='softmax')])\n",
    "\n",
    "model.compile(optimizer='adam',\n",
    "              loss='categorical_crossentropy',\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## One-hot кодировка разметки"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_labels = tf.keras.utils.to_categorical(y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(y_train.shape)\n",
    "print(y_train_labels.shape)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Запускаем обучение"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "model.fit(x_train, \n",
    "          y_train_labels,\n",
    "          batch_size=32, \n",
    "          epochs=5,\n",
    "          validation_split=0.2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Предсказываем класс объекта"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred = model.predict(x_val)\n",
    "y_pred_labels = np.argmax(y_pred, axis=1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Оцениваем качество решение на валидационной выборке"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import accuracy_score\n",
    "print('Accuracy: %s' % accuracy_score(y_val, y_pred_labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import classification_report\n",
    "print(classification_report(y_val, y_pred_labels))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.metrics import confusion_matrix\n",
    "cm = confusion_matrix(y_val, y_pred_labels)\n",
    "\n",
    "plt.figure(figsize=(10, 8))\n",
    "plt.imshow(cm, cmap=plt.cm.Blues)\n",
    "plt.xlabel(\"Predicted labels\")\n",
    "plt.ylabel(\"True labels\")\n",
    "plt.xticks([], [])\n",
    "plt.yticks([], [])\n",
    "plt.title('Confusion matrix ')\n",
    "plt.colorbar()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Визуализируем предсказания для тестовой выборки"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "y_pred_test = model.predict(test_img)\n",
    "y_pred_test_labels = np.argmax(y_pred_test, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "fig = plt.figure(figsize=(20, 10))\n",
    "indices = np.random.choice(range(len(test_img)), 5)\n",
    "img_prediction = zip(test_img[indices], y_pred_test_labels[indices])\n",
    "for i, (img, pred) in enumerate(img_prediction, 1):\n",
    "    subplot = fig.add_subplot(1, 5, i)\n",
    "    plt.imshow(img[...,0], cmap='gray');\n",
    "    subplot.set_title('%d' % pred);"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Готовим файл для отправки"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "with open('submit.txt', 'w') as dst:\n",
    "    dst.write('ImageId,Label\\n')\n",
    "    for i, p in enumerate(y_pred_test_labels, 1):\n",
    "        dst.write('%s,%d\\n' % (i, p))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Your submission scored 0.96814"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}