upload ch09 files.

09_rnn_tensorflow.ipynb

@@ -0,0 +1,289 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# 循环神经网络—TensorFlow"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from tensorflow.examples.tutorials.mnist import input_data\n",
+    "import tensorflow as tf\n",
+    "import matplotlib.pyplot as plt"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 导入 MNIST 数据集"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "WARNING:tensorflow:From <ipython-input-2-3f5d8b810895>:1: read_data_sets (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\n",
+      "Instructions for updating:\n",
+      "Please use alternatives such as official/mnist/dataset.py from tensorflow/models.\n",
+      "WARNING:tensorflow:From C:\\ProgramData\\Anaconda3\\envs\\tensorflow\\lib\\site-packages\\tensorflow\\contrib\\learn\\python\\learn\\datasets\\mnist.py:260: maybe_download (from tensorflow.contrib.learn.python.learn.datasets.base) is deprecated and will be removed in a future version.\n",
+      "Instructions for updating:\n",
+      "Please write your own downloading logic.\n",
+      "WARNING:tensorflow:From C:\\ProgramData\\Anaconda3\\envs\\tensorflow\\lib\\site-packages\\tensorflow\\contrib\\learn\\python\\learn\\datasets\\mnist.py:262: extract_images (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\n",
+      "Instructions for updating:\n",
+      "Please use tf.data to implement this functionality.\n",
+      "Extracting ./datasets/ch08/tensorflow/MNIST\\train-images-idx3-ubyte.gz\n",
+      "WARNING:tensorflow:From C:\\ProgramData\\Anaconda3\\envs\\tensorflow\\lib\\site-packages\\tensorflow\\contrib\\learn\\python\\learn\\datasets\\mnist.py:267: extract_labels (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\n",
+      "Instructions for updating:\n",
+      "Please use tf.data to implement this functionality.\n",
+      "Extracting ./datasets/ch08/tensorflow/MNIST\\train-labels-idx1-ubyte.gz\n",
+      "WARNING:tensorflow:From C:\\ProgramData\\Anaconda3\\envs\\tensorflow\\lib\\site-packages\\tensorflow\\contrib\\learn\\python\\learn\\datasets\\mnist.py:110: dense_to_one_hot (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\n",
+      "Instructions for updating:\n",
+      "Please use tf.one_hot on tensors.\n",
+      "Extracting ./datasets/ch08/tensorflow/MNIST\\t10k-images-idx3-ubyte.gz\n",
+      "Extracting ./datasets/ch08/tensorflow/MNIST\\t10k-labels-idx1-ubyte.gz\n",
+      "WARNING:tensorflow:From C:\\ProgramData\\Anaconda3\\envs\\tensorflow\\lib\\site-packages\\tensorflow\\contrib\\learn\\python\\learn\\datasets\\mnist.py:290: DataSet.__init__ (from tensorflow.contrib.learn.python.learn.datasets.mnist) is deprecated and will be removed in a future version.\n",
+      "Instructions for updating:\n",
+      "Please use alternatives such as official/mnist/dataset.py from tensorflow/models.\n"
+     ]
+    }
+   ],
+   "source": [
+    "mnist = input_data.read_data_sets('./datasets/ch08/tensorflow/MNIST',one_hot=True)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 3,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "(55000, 784)\n",
+      "(55000, 10)\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(mnist.train.images.shape)\n",
+    "print(mnist.train.labels.shape)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "(5000, 784)\n",
+      "(5000, 10)\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(mnist.validation.images.shape)\n",
+    "print(mnist.validation.labels.shape)"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "(10000, 784)\n",
+      "(10000, 10)\n"
+     ]
+    }
+   ],
+   "source": [
+    "print(mnist.test.images.shape)\n",
+    "print(mnist.test.labels.shape)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 定义占位符 placeholder"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 6,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "# 参数设置\n",
+    "batch_size = 100        # BATCH 的大小，相当于一次处理100个image\n",
+    "time_step = 28          # 一个LSTM中，输入序列的长度Tx，image有28行\n",
+    "input_size = 28         # 单个x向量长度，image有28列\n",
+    "lr = 0.001              # 学习率\n",
+    "num_units = 100         # 隐藏层多少个LTSM单元\n",
+    "iterations =1000        # 迭代训练次数\n",
+    "classes =10             # 输出大小，0-9十个数字的概率\n",
+    "\n",
+    "# 定义 placeholders\n",
+    "# 维度是[batch_size，time_step * input_size]\n",
+    "x = tf.placeholder(tf.float32, [None, time_step * input_size]) \n",
+    "# 输入的是二维数据，将其还原为三维，维度是[batch_size, time_step, input_size]\n",
+    "x_image = tf.reshape(x, [-1, time_step, input_size])                   \n",
+    "y = tf.placeholder(tf.int32, [None, classes])                     "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 定义RNN（LSTM）结构"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 7,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "rnn_cell = tf.contrib.rnn.BasicLSTMCell(num_units=num_units) \n",
+    "outputs,final_state = tf.nn.dynamic_rnn(\n",
+    "    cell=rnn_cell,              # 选择传入的cell\n",
+    "    inputs=x_image,             # 传入的数据\n",
+    "    initial_state=None,         # 初始状态\n",
+    "    dtype=tf.float32,           # 数据类型\n",
+    "    time_major=False,           # False: (batch, time_step, input); True: (time_step, batch, input)，这里根据x_image结构选择False\n",
+    ")\n",
+    "output = tf.layers.dense(inputs=outputs[:, -1, :], units=classes)    "
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 定义损失函数与优化算法"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 8,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "cross_entropy = tf.losses.softmax_cross_entropy(onehot_labels=y, logits=output)      # 计算loss\n",
+    "train_step = tf.train.AdamOptimizer(lr).minimize(cross_entropy)      #选择优化方法"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 9,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "#判断预测标签和实际标签是否匹配\n",
+    "correct_prediction = tf.equal(tf.argmax(y, axis=1),tf.argmax(output, axis=1))\n",
+    "accuracy = tf.reduce_mean(tf.cast(correct_prediction,'float'))  #计算正确率"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## 训练并验证准确率"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "train accuracy 0.560\n",
+      "train accuracy 0.740\n",
+      "train accuracy 0.900\n",
+      "train accuracy 0.810\n",
+      "train accuracy 0.940\n",
+      "train accuracy 0.890\n",
+      "train accuracy 0.880\n",
+      "train accuracy 0.910\n",
+      "train accuracy 0.910\n",
+      "train accuracy 0.930\n",
+      "train accuracy 0.870\n",
+      "train accuracy 0.960\n",
+      "train accuracy 0.960\n",
+      "train accuracy 0.940\n",
+      "train accuracy 0.930\n",
+      "train accuracy 0.960\n",
+      "train accuracy 0.940\n",
+      "train accuracy 0.980\n",
+      "train accuracy 0.920\n",
+      "train accuracy 0.980\n",
+      "test accuracy 0.958\n"
+     ]
+    }
+   ],
+   "source": [
+    "sess = tf.Session()\n",
+    "init = tf.global_variables_initializer()\n",
+    "sess.run(init)\n",
+    "\n",
+    "for i in range(iterations):\n",
+    "    batch_x, batch_y = mnist.train.next_batch(batch_size)\n",
+    "    sess.run(train_step, feed_dict={x: batch_x, y: batch_y})\n",
+    "    if (i+1) % 50 == 0:\n",
+    "        print(\"train accuracy %.3f\" % accuracy.eval(session = sess,\n",
+    "                feed_dict = {x:batch_x, y:batch_y}))\n",
+    "print(\"test accuracy %.3f\" % accuracy.eval(session = sess,\n",
+    "    feed_dict = {x:mnist.test.images, y:mnist.test.labels}))"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "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": "3.5.6"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

README.md

@@ -14,4 +14,8 @@
 
 - ## [08_cnn_pytorch.ipynb](https://github.com/RedstoneWill/dl-from-scratch/blob/master/08_cnn_pytorch.ipynb)
 
-- ## [08_cnn_tensorflow.ipynb](https://github.com/RedstoneWill/dl-from-scratch/blob/master/08_cnn_tensorflow.ipynb)
+- ## [08_cnn_tensorflow.ipynb](https://github.com/RedstoneWill/dl-from-scratch/blob/master/08_cnn_tensorflow.ipynb)
+
+- ## [09_rnn_pytorch.ipynb](https://github.com/RedstoneWill/dl-from-scratch/blob/master/09_rnn_pytorch.ipynb)
+
+- ## [09_rnn_tensorflow.ipynb](https://github.com/RedstoneWill/dl-from-scratch/blob/master/09_rnn_tensorflow.ipynb)