03_CNN_basics/cnn_tf.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# TensorFlow卷积神经网络(CNN)示例 - 高级API\n",
    "### Convolutional Neural Network Example - tf.layers API"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## CNN网络结构图示\n",
    "\n",
    "![CNN](http://personal.ie.cuhk.edu.hk/~ccloy/project_target_code/images/fig3.png)\n",
    "\n",
    "## MNIST数据库\n",
    "\n",
    "![MNIST Dataset](http://neuralnetworksanddeeplearning.com/images/mnist_100_digits.png)\n",
    "\n",
    "More info: http://yann.lecun.com/exdb/mnist/"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Extracting /tmp/data/train-images-idx3-ubyte.gz\n",
      "Extracting /tmp/data/train-labels-idx1-ubyte.gz\n",
      "Extracting /tmp/data/t10k-images-idx3-ubyte.gz\n",
      "Extracting /tmp/data/t10k-labels-idx1-ubyte.gz\n"
     ]
    }
   ],
   "source": [
    "from __future__ import division, print_function, absolute_import\n",
    "\n",
    "# Import MNIST data，MNIST数据集导入\n",
    "from tensorflow.examples.tutorials.mnist import input_data\n",
    "mnist = input_data.read_data_sets(\"/tmp/data/\", one_hot=False)\n",
    "\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Training Parameters，超参数\n",
    "learning_rate = 0.001 #学习率\n",
    "num_steps = 2000 # 训练步数\n",
    "batch_size = 128 # 训练数据批的大小\n",
    "\n",
    "# Network Parameters，网络参数\n",
    "num_input = 784 # MNIST数据输入 (img shape: 28*28)\n",
    "num_classes = 10 # MNIST所有类别 (0-9 digits)\n",
    "dropout = 0.75 # Dropout, probability to keep units = (1-p)，保留神经元相应的概率为(1-p)=(1-0.75)=0.25"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# Create the neural network，创建深度神经网络\n",
    "def conv_net(x_dict, n_classes, dropout, reuse, is_training):\n",
    "    \n",
    "    # Define a scope for reusing the variables，确定命名空间\n",
    "    with tf.variable_scope('ConvNet', reuse=reuse):\n",
    "        # TF Estimator类型的输入为像素\n",
    "        x = x_dict['images']\n",
    "\n",
    "        # MNIST数据输入格式为一位向量，包含784个特征 (28*28像素)\n",
    "        # 用reshape函数改变形状以匹配图像的尺寸 [高 x 宽 x 通道数]\n",
    "        # 输入张量的尺度为四维: [(每一)批数据的数目, 高，宽，通道数]\n",
    "        x = tf.reshape(x, shape=[-1, 28, 28, 1])\n",
    "\n",
    "        # 卷积层，32个卷积核，尺寸为5x5\n",
    "        conv1 = tf.layers.conv2d(x, 32, 5, activation=tf.nn.relu)\n",
    "        # 最大池化层，步长为2，无需学习任何参量\n",
    "        conv1 = tf.layers.max_pooling2d(conv1, 2, 2)\n",
    "\n",
    "        # 卷积层，64个卷积核，尺寸为3x3\n",
    "        conv2 = tf.layers.conv2d(conv1, 64, 3, activation=tf.nn.relu)\n",
    "        # 最大池化层，步长为2，无需学习任何参量\n",
    "        conv2 = tf.layers.max_pooling2d(conv2, 2, 2)\n",
    "\n",
    "        # 展开特征为一维向量，以输入全连接层\n",
    "        fc1 = tf.contrib.layers.flatten(conv2)\n",
    "\n",
    "        # 全连接层\n",
    "        fc1 = tf.layers.dense(fc1, 1024)\n",
    "        # 应用Dropout (训练时打开，测试时关闭)\n",
    "        fc1 = tf.layers.dropout(fc1, rate=dropout, training=is_training)\n",
    "\n",
    "        # 输出层，预测类别\n",
    "        out = tf.layers.dense(fc1, n_classes)\n",
    "\n",
    "    return out"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 确定模型功能 (参照TF Estimator模版)\n",
    "def model_fn(features, labels, mode):\n",
    "    \n",
    "    # 构建神经网络\n",
    "    # 因为dropout在训练与测试时的特性不一，我们此处为训练和测试过程创建两个独立但共享权值的计算图\n",
    "    logits_train = conv_net(features, num_classes, dropout, reuse=False, is_training=True)\n",
    "    logits_test = conv_net(features, num_classes, dropout, reuse=True, is_training=False)\n",
    "    \n",
    "    # 预测\n",
    "    pred_classes = tf.argmax(logits_test, axis=1)\n",
    "    pred_probas = tf.nn.softmax(logits_test)\n",
    "    \n",
    "    if mode == tf.estimator.ModeKeys.PREDICT:\n",
    "        return tf.estimator.EstimatorSpec(mode, predictions=pred_classes) \n",
    "        \n",
    "    # 确定误差函数与优化器\n",
    "    loss_op = tf.reduce_mean(tf.nn.sparse_softmax_cross_entropy_with_logits(\n",
    "        logits=logits_train, labels=tf.cast(labels, dtype=tf.int32)))\n",
    "    optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate)\n",
    "    train_op = optimizer.minimize(loss_op, global_step=tf.train.get_global_step())\n",
    "    \n",
    "    # 评估模型精确度\n",
    "    acc_op = tf.metrics.accuracy(labels=labels, predictions=pred_classes)\n",
    "    \n",
    "    # TF Estimators需要返回EstimatorSpec\n",
    "    estim_specs = tf.estimator.EstimatorSpec(\n",
    "      mode=mode,\n",
    "      predictions=pred_classes,\n",
    "      loss=loss_op,\n",
    "      train_op=train_op,\n",
    "      eval_metric_ops={'accuracy': acc_op})\n",
    "\n",
    "    return estim_specs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Using default config.\n",
      "WARNING:tensorflow:Using temporary folder as model directory: /tmp/tmpdhd6F4\n",
      "INFO:tensorflow:Using config: {'_save_checkpoints_secs': 600, '_session_config': None, '_keep_checkpoint_max': 5, '_tf_random_seed': 1, '_keep_checkpoint_every_n_hours': 10000, '_log_step_count_steps': 100, '_save_checkpoints_steps': None, '_model_dir': '/tmp/tmpdhd6F4', '_save_summary_steps': 100}\n"
     ]
    }
   ],
   "source": [
    "# 构建Estimator\n",
    "model = tf.estimator.Estimator(model_fn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Create CheckpointSaverHook.\n",
      "INFO:tensorflow:Saving checkpoints for 1 into /tmp/tmpdhd6F4/model.ckpt.\n",
      "INFO:tensorflow:loss = 2.39026, step = 1\n",
      "INFO:tensorflow:global_step/sec: 238.314\n",
      "INFO:tensorflow:loss = 0.237997, step = 101 (0.421 sec)\n",
      "INFO:tensorflow:global_step/sec: 255.312\n",
      "INFO:tensorflow:loss = 0.0954537, step = 201 (0.392 sec)\n",
      "INFO:tensorflow:global_step/sec: 257.194\n",
      "INFO:tensorflow:loss = 0.121477, step = 301 (0.389 sec)\n",
      "INFO:tensorflow:global_step/sec: 255.018\n",
      "INFO:tensorflow:loss = 0.0539927, step = 401 (0.392 sec)\n",
      "INFO:tensorflow:global_step/sec: 254.293\n",
      "INFO:tensorflow:loss = 0.0440369, step = 501 (0.393 sec)\n",
      "INFO:tensorflow:global_step/sec: 256.501\n",
      "INFO:tensorflow:loss = 0.0247431, step = 601 (0.390 sec)\n",
      "INFO:tensorflow:global_step/sec: 252.956\n",
      "INFO:tensorflow:loss = 0.0738082, step = 701 (0.395 sec)\n",
      "INFO:tensorflow:global_step/sec: 253.222\n",
      "INFO:tensorflow:loss = 0.134998, step = 801 (0.395 sec)\n",
      "INFO:tensorflow:global_step/sec: 255.606\n",
      "INFO:tensorflow:loss = 0.00438448, step = 901 (0.391 sec)\n",
      "INFO:tensorflow:global_step/sec: 256.306\n",
      "INFO:tensorflow:loss = 0.0471991, step = 1001 (0.390 sec)\n",
      "INFO:tensorflow:global_step/sec: 255.352\n",
      "INFO:tensorflow:loss = 0.0371172, step = 1101 (0.392 sec)\n",
      "INFO:tensorflow:global_step/sec: 253.277\n",
      "INFO:tensorflow:loss = 0.0129522, step = 1201 (0.395 sec)\n",
      "INFO:tensorflow:global_step/sec: 252.49\n",
      "INFO:tensorflow:loss = 0.039862, step = 1301 (0.396 sec)\n",
      "INFO:tensorflow:global_step/sec: 253.902\n",
      "INFO:tensorflow:loss = 0.0520571, step = 1401 (0.394 sec)\n",
      "INFO:tensorflow:global_step/sec: 255.572\n",
      "INFO:tensorflow:loss = 0.0307549, step = 1501 (0.392 sec)\n",
      "INFO:tensorflow:global_step/sec: 254.32\n",
      "INFO:tensorflow:loss = 0.0108862, step = 1601 (0.393 sec)\n",
      "INFO:tensorflow:global_step/sec: 255.62\n",
      "INFO:tensorflow:loss = 0.0294434, step = 1701 (0.391 sec)\n",
      "INFO:tensorflow:global_step/sec: 254.349\n",
      "INFO:tensorflow:loss = 0.0179781, step = 1801 (0.393 sec)\n",
      "INFO:tensorflow:global_step/sec: 255.508\n",
      "INFO:tensorflow:loss = 0.0375271, step = 1901 (0.391 sec)\n",
      "INFO:tensorflow:Saving checkpoints for 2000 into /tmp/tmpdhd6F4/model.ckpt.\n",
      "INFO:tensorflow:Loss for final step: 0.00440777.\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<tensorflow.python.estimator.estimator.Estimator at 0x7fb80ca55c90>"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 确定训练输入函数\n",
    "input_fn = tf.estimator.inputs.numpy_input_fn(\n",
    "    x={'images': mnist.train.images}, y=mnist.train.labels,\n",
    "    batch_size=batch_size, num_epochs=None, shuffle=True)\n",
    "# 开始训练模型\n",
    "model.train(input_fn, steps=num_steps)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Starting evaluation at 2017-08-21-14:25:29\n",
      "INFO:tensorflow:Restoring parameters from /tmp/tmpdhd6F4/model.ckpt-2000\n",
      "INFO:tensorflow:Finished evaluation at 2017-08-21-14:25:29\n",
      "INFO:tensorflow:Saving dict for global step 2000: accuracy = 0.9908, global_step = 2000, loss = 0.0382241\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "{'accuracy': 0.99080002, 'global_step': 2000, 'loss': 0.038224086}"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# 评判模型\n",
    "# 确定评判用输入函数\n",
    "input_fn = tf.estimator.inputs.numpy_input_fn(\n",
    "    x={'images': mnist.test.images}, y=mnist.test.labels,\n",
    "    batch_size=batch_size, shuffle=False)\n",
    "model.evaluate(input_fn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INFO:tensorflow:Restoring parameters from /tmp/tmpdhd6F4/model.ckpt-2000\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nkiL8QFKEH0iK8ANJEX4gKcIPJPVP82g/p9/JjhUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb80c478dd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model prediction: 7\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb80c489b50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model prediction: 2\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb80c6b4e10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model prediction: 1\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAP8AAAD8CAYAAAC4nHJkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nZZUAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x7fb80c3af290>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model prediction: 0\n"
     ]
    }
   ],
   "source": [
    "# 预测单个图像\n",
    "n_images = 4\n",
    "# 从数据集得到测试图像\n",
    "test_images = mnist.test.images[:n_images]\n",
    "# 准备输入数据\n",
    "input_fn = tf.estimator.inputs.numpy_input_fn(\n",
    "    x={'images': test_images}, shuffle=False)\n",
    "# 用训练好的模型预测图片类别\n",
    "preds = list(model.predict(input_fn))\n",
    "\n",
    "# 可视化显示\n",
    "for i in range(n_images):\n",
    "    plt.imshow(np.reshape(test_images[i], [28, 28]), cmap='gray')\n",
    "    plt.show()\n",
    "    print(\"Model prediction:\", preds[i])"
   ]
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "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": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}