5-3 Tensorboard 网络运行-可视化网络参数-可视化调参.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From <ipython-input-1-5e1a66207665>:4: 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 D:\\SoftWare\\Anaconda3\\envs\\pyDL\\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 D:\\SoftWare\\Anaconda3\\envs\\pyDL\\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 MNIST_data\\train-images-idx3-ubyte.gz\n",
      "WARNING:tensorflow:From D:\\SoftWare\\Anaconda3\\envs\\pyDL\\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 MNIST_data\\train-labels-idx1-ubyte.gz\n",
      "WARNING:tensorflow:From D:\\SoftWare\\Anaconda3\\envs\\pyDL\\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 MNIST_data\\t10k-images-idx3-ubyte.gz\n",
      "Extracting MNIST_data\\t10k-labels-idx1-ubyte.gz\n",
      "WARNING:tensorflow:From D:\\SoftWare\\Anaconda3\\envs\\pyDL\\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": [
    "import tensorflow as tf\n",
    "from tensorflow.examples.tutorials.mnist import input_data\n",
    "# 载入数据\n",
    "mnist = input_data.read_data_sets('MNIST_data', one_hot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Iter 0 Testing Accuracy: 0.7957\n",
      "Iter 1 Testing Accuracy: 0.8618\n",
      "Iter 2 Testing Accuracy: 0.8757\n",
      "Iter 3 Testing Accuracy: 0.8821\n",
      "Iter 4 Testing Accuracy: 0.8869\n",
      "Iter 5 Testing Accuracy: 0.8933\n",
      "Iter 6 Testing Accuracy: 0.8948\n",
      "Iter 7 Testing Accuracy: 0.8977\n",
      "Iter 8 Testing Accuracy: 0.8999\n",
      "Iter 9 Testing Accuracy: 0.9016\n",
      "Iter 10 Testing Accuracy: 0.9036\n",
      "Iter 11 Testing Accuracy: 0.9037\n",
      "Iter 12 Testing Accuracy: 0.9053\n",
      "Iter 13 Testing Accuracy: 0.9055\n",
      "Iter 14 Testing Accuracy: 0.9073\n",
      "Iter 15 Testing Accuracy: 0.9076\n",
      "Iter 16 Testing Accuracy: 0.9089\n",
      "Iter 17 Testing Accuracy: 0.9091\n",
      "Iter 18 Testing Accuracy: 0.9105\n",
      "Iter 19 Testing Accuracy: 0.91\n",
      "Iter 20 Testing Accuracy: 0.9116\n",
      "Iter 21 Testing Accuracy: 0.9123\n",
      "Iter 22 Testing Accuracy: 0.9118\n",
      "Iter 23 Testing Accuracy: 0.9126\n",
      "Iter 24 Testing Accuracy: 0.9128\n",
      "Iter 25 Testing Accuracy: 0.9134\n",
      "Iter 26 Testing Accuracy: 0.9138\n",
      "Iter 27 Testing Accuracy: 0.9142\n",
      "Iter 28 Testing Accuracy: 0.9147\n",
      "Iter 29 Testing Accuracy: 0.915\n",
      "Iter 30 Testing Accuracy: 0.9149\n",
      "Iter 31 Testing Accuracy: 0.916\n",
      "Iter 32 Testing Accuracy: 0.916\n",
      "Iter 33 Testing Accuracy: 0.9164\n",
      "Iter 34 Testing Accuracy: 0.9171\n",
      "Iter 35 Testing Accuracy: 0.9176\n",
      "Iter 36 Testing Accuracy: 0.9178\n",
      "Iter 37 Testing Accuracy: 0.9176\n",
      "Iter 38 Testing Accuracy: 0.9172\n",
      "Iter 39 Testing Accuracy: 0.9182\n",
      "Iter 40 Testing Accuracy: 0.9183\n",
      "Iter 41 Testing Accuracy: 0.9182\n",
      "Iter 42 Testing Accuracy: 0.9179\n",
      "Iter 43 Testing Accuracy: 0.9187\n",
      "Iter 44 Testing Accuracy: 0.9185\n",
      "Iter 45 Testing Accuracy: 0.9188\n",
      "Iter 46 Testing Accuracy: 0.9184\n",
      "Iter 47 Testing Accuracy: 0.9185\n",
      "Iter 48 Testing Accuracy: 0.9186\n",
      "Iter 49 Testing Accuracy: 0.9185\n",
      "Iter 50 Testing Accuracy: 0.9202\n"
     ]
    }
   ],
   "source": [
    "# 批次的大小\n",
    "batch_size = 128\n",
    "n_batch = mnist.train.num_examples // batch_size\n",
    "\n",
    "# 参数概要\n",
    "def variable_summaries(var):\n",
    "    with tf.name_scope('summaries'):\n",
    "        mean = tf.reduce_mean(var)\n",
    "        tf.summary.scalar('mean',mean) # 平均值\n",
    "        with tf.name_scope('stddev'):\n",
    "            stddev = tf.sqrt(tf.reduce_mean(tf.square(var - mean)))\n",
    "        tf.summary.scalar('stddev',stddev) # 标准差\n",
    "        tf.summary.scalar('max', tf.reduce_max(var)) # 最大值\n",
    "        tf.summary.scalar('min', tf.reduce_min(var)) # 最小值\n",
    "        tf.summary.histogram('histogram',var) # 直方图\n",
    "\n",
    "\n",
    "# 命名空间\n",
    "with tf.name_scope('input'):\n",
    "    x = tf.placeholder(tf.float32, [None,784],name=\"X-input\")\n",
    "    y = tf.placeholder(tf.float32, [None, 10],name='y-input')\n",
    "\n",
    "# 创建一个简单的神经网络\n",
    "with tf.name_scope('layer'):\n",
    "    with tf.name_scope('weights'):\n",
    "        W = tf.Variable(tf.zeros([784,10]),name='W')\n",
    "        variable_summaries(W)\n",
    "    with tf.name_scope('biases'):\n",
    "        b = tf.Variable(tf.zeros([1, 10]),name='b')\n",
    "        variable_summaries(b)\n",
    "    with tf.name_scope('xw_plus_b'):\n",
    "        wx_plus_b = tf.matmul(x,W) + b\n",
    "    with tf.name_scope('softmax'):\n",
    "        prediction = tf.nn.softmax(wx_plus_b)\n",
    "\n",
    "# 代价函数\n",
    "with tf.name_scope('loss'):\n",
    "    loss = tf.reduce_mean(tf.square(y-prediction))\n",
    "    tf.summary.scalar('loss', loss)\n",
    "# 梯度下降法\n",
    "with tf.name_scope('train'):\n",
    "    train_step = tf.train.GradientDescentOptimizer(0.2).minimize(loss)\n",
    "    \n",
    "# 初始化变量\n",
    "init = tf.global_variables_initializer()\n",
    "\n",
    "# 得到一个布尔型列表，存放结果是否正确\n",
    "with tf.name_scope('accuracy'):\n",
    "    with tf.name_scope('correct_prediction'):\n",
    "        correct_prediction = tf.equal(tf.argmax(y,1), tf.argmax(prediction,1)) #argmax 返回一维张量中最大值索引\n",
    "    # 求准确率\n",
    "    with tf.name_scope('accuracy'):\n",
    "        accuracy = tf.reduce_mean(tf.cast(correct_prediction,tf.float32)) # 把布尔值转换为浮点型求平均数\n",
    "        tf.summary.scalar('accuracy', accuracy)\n",
    "        \n",
    "# 合并所有summary\n",
    "merged = tf.summary.merge_all()\n",
    "        \n",
    "with tf.Session() as sess:\n",
    "    sess.run(init)\n",
    "    writer = tf.summary.FileWriter('logs/',sess.graph)\n",
    "    for epoch in range(51):\n",
    "        for batch in range(n_batch):\n",
    "            # 获得批次数据\n",
    "            batch_xs, batch_ys = mnist.train.next_batch(batch_size)\n",
    "            # 运行网络并记录log\n",
    "            summary,_ = sess.run([merged, train_step], feed_dict={x:batch_xs, y:batch_ys})\n",
    "        # 记录变量\n",
    "        writer.add_summary(summary,epoch)\n",
    "        acc = sess.run(accuracy, feed_dict={x:mnist.test.images,y:mnist.test.labels})\n",
    "        print(\"Iter \" + str(epoch) + \" Testing Accuracy: \" + str(acc))\n",
    "    writer.close()\n",
    "    writer.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# 补充资料\n",
    "本节所讲的记录网络数据的内容代码比较简单，笔者从网上找到一份更优美的代码，\n",
    "[Tensorflow的可视化工具Tensorboard的初步使用](https://blog.csdn.net/sinat_33761963/article/category/6564135)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From <ipython-input-1-5e1a66207665>:4: 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 D:\\SoftWare\\Anaconda3\\envs\\pyDL\\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 D:\\SoftWare\\Anaconda3\\envs\\pyDL\\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 MNIST_data\\train-images-idx3-ubyte.gz\n",
      "WARNING:tensorflow:From D:\\SoftWare\\Anaconda3\\envs\\pyDL\\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 MNIST_data\\train-labels-idx1-ubyte.gz\n",
      "WARNING:tensorflow:From D:\\SoftWare\\Anaconda3\\envs\\pyDL\\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 MNIST_data\\t10k-images-idx3-ubyte.gz\n",
      "Extracting MNIST_data\\t10k-labels-idx1-ubyte.gz\n",
      "WARNING:tensorflow:From D:\\SoftWare\\Anaconda3\\envs\\pyDL\\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": [
    "import tensorflow as tf\n",
    "from tensorflow.examples.tutorials.mnist import input_data\n",
    "# 载入数据\n",
    "mnist = input_data.read_data_sets('MNIST_data', one_hot=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "max_step = 1000\n",
    "learning_rate = 0.001\n",
    "dropout = 0.9\n",
    "\n",
    "log_dir = \"Logs/log-5.3\"\n",
    "\n",
    "with tf.name_scope('input'):\n",
    "    x = tf.placeholder(tf.float32, [None, 784], name='x-input')\n",
    "    y_ = tf.placeholder(tf.float32, [None, 10], name='y-input')\n",
    "    keep_prob = tf.placeholder(tf.float32)\n",
    "    \n",
    "with tf.name_scope('input_reshape'):\n",
    "    image_shaped_input = tf.reshape(x, [-1, 28, 28, 1])\n",
    "    tf.summary.image('input', image_shaped_input, 10)\n",
    "    \n",
    "def weight_variable(shape):\n",
    "    return tf.Variable(tf.truncated_normal(shape, stddev=0.1))\n",
    "\n",
    "def bias_variable(shape):\n",
    "    return tf.Variable(tf.constant(0.1, shape=shape))\n",
    "\n",
    "def variable_summaries(var):\n",
    "    with tf.name_scope('summaries'):\n",
    "        mean = tf.reduce_mean(var)\n",
    "        tf.summary.scalar('mean', mean)\n",
    "        \n",
    "        with tf.name_scope('stddev'):\n",
    "            stddev = tf.sqrt(tf.reduce_mean(tf.square(var-mean)))\n",
    "        tf.summary.scalar('stddev', stddev)\n",
    "        tf.summary.scalar('max', tf.reduce_max(var))\n",
    "        tf.summary.scalar('min', tf.reduce_min(var))\n",
    "        tf.summary.histogram('histogram', var)\n",
    "        \n",
    "def nn_layer(input_tensor, input_dim, output_dim, layer_name, act=tf.nn.relu):\n",
    "    with tf.name_scope(layer_name):\n",
    "        with tf.name_scope('weights'):\n",
    "            weights = weight_variable([input_dim, output_dim])\n",
    "            variable_summaries(weights)\n",
    "        with tf.name_scope('biases'):\n",
    "            biases = bias_variable([output_dim])\n",
    "            variable_summaries(biases)\n",
    "        with tf.name_scope('linear_compute'):\n",
    "            preactivate = tf.matmul(input_tensor, weights) + biases\n",
    "            tf.summary.histogram('linear', preactivate)\n",
    "        activations = act(preactivate, name='activation')\n",
    "        tf.summary.histogram('activations', activations)\n",
    "        \n",
    "        return activations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "WARNING:tensorflow:From <ipython-input-3-0d0774accdbe>:13: softmax_cross_entropy_with_logits (from tensorflow.python.ops.nn_ops) is deprecated and will be removed in a future version.\n",
      "Instructions for updating:\n",
      "\n",
      "Future major versions of TensorFlow will allow gradients to flow\n",
      "into the labels input on backprop by default.\n",
      "\n",
      "See @{tf.nn.softmax_cross_entropy_with_logits_v2}.\n",
      "\n",
      "Accuracy at step 0: 0.0844\n",
      "Accuracy at step 10: 0.7245\n",
      "Accuracy at step 20: 0.8235\n",
      "Accuracy at step 30: 0.8616\n",
      "Accuracy at step 40: 0.8854\n",
      "Accuracy at step 50: 0.8961\n",
      "Accuracy at step 60: 0.899\n",
      "Accuracy at step 70: 0.906\n",
      "Accuracy at step 80: 0.9112\n",
      "Accuracy at step 90: 0.9152\n",
      "Accuracy at step 100: 0.9166\n",
      "Accuracy at step 110: 0.9126\n",
      "Accuracy at step 120: 0.9236\n",
      "Accuracy at step 130: 0.9288\n",
      "Accuracy at step 140: 0.9261\n",
      "Accuracy at step 150: 0.9275\n",
      "Accuracy at step 160: 0.9297\n",
      "Accuracy at step 170: 0.9327\n",
      "Accuracy at step 180: 0.9339\n",
      "Accuracy at step 190: 0.9329\n",
      "Accuracy at step 200: 0.9342\n",
      "Accuracy at step 210: 0.9357\n",
      "Accuracy at step 220: 0.9317\n",
      "Accuracy at step 230: 0.9358\n",
      "Accuracy at step 240: 0.9414\n",
      "Accuracy at step 250: 0.9407\n",
      "Accuracy at step 260: 0.9373\n",
      "Accuracy at step 270: 0.9449\n",
      "Accuracy at step 280: 0.9441\n",
      "Accuracy at step 290: 0.9407\n",
      "Accuracy at step 300: 0.9448\n",
      "Accuracy at step 310: 0.9416\n",
      "Accuracy at step 320: 0.9492\n",
      "Accuracy at step 330: 0.9465\n",
      "Accuracy at step 340: 0.9413\n",
      "Accuracy at step 350: 0.9484\n",
      "Accuracy at step 360: 0.9512\n",
      "Accuracy at step 370: 0.9498\n",
      "Accuracy at step 380: 0.9505\n",
      "Accuracy at step 390: 0.9501\n",
      "Accuracy at step 400: 0.9511\n",
      "Accuracy at step 410: 0.9521\n",
      "Accuracy at step 420: 0.9528\n",
      "Accuracy at step 430: 0.9533\n",
      "Accuracy at step 440: 0.9565\n",
      "Accuracy at step 450: 0.9508\n",
      "Accuracy at step 460: 0.9584\n",
      "Accuracy at step 470: 0.958\n",
      "Accuracy at step 480: 0.955\n",
      "Accuracy at step 490: 0.9562\n",
      "Accuracy at step 500: 0.9566\n",
      "Accuracy at step 510: 0.9562\n",
      "Accuracy at step 520: 0.9576\n",
      "Accuracy at step 530: 0.9559\n",
      "Accuracy at step 540: 0.9582\n",
      "Accuracy at step 550: 0.9552\n",
      "Accuracy at step 560: 0.9581\n",
      "Accuracy at step 570: 0.9598\n",
      "Accuracy at step 580: 0.9596\n",
      "Accuracy at step 590: 0.9595\n",
      "Accuracy at step 600: 0.9567\n",
      "Accuracy at step 610: 0.9613\n",
      "Accuracy at step 620: 0.96\n",
      "Accuracy at step 630: 0.9612\n",
      "Accuracy at step 640: 0.9619\n",
      "Accuracy at step 650: 0.9635\n",
      "Accuracy at step 660: 0.9607\n",
      "Accuracy at step 670: 0.9598\n",
      "Accuracy at step 680: 0.9637\n",
      "Accuracy at step 690: 0.9621\n",
      "Accuracy at step 700: 0.9634\n",
      "Accuracy at step 710: 0.9642\n",
      "Accuracy at step 720: 0.9649\n",
      "Accuracy at step 730: 0.9644\n",
      "Accuracy at step 740: 0.9657\n",
      "Accuracy at step 750: 0.9639\n",
      "Accuracy at step 760: 0.9666\n",
      "Accuracy at step 770: 0.9657\n",
      "Accuracy at step 780: 0.9669\n",
      "Accuracy at step 790: 0.9653\n",
      "Accuracy at step 800: 0.9671\n",
      "Accuracy at step 810: 0.966\n",
      "Accuracy at step 820: 0.9658\n",
      "Accuracy at step 830: 0.9655\n",
      "Accuracy at step 840: 0.967\n",
      "Accuracy at step 850: 0.9666\n",
      "Accuracy at step 860: 0.9653\n",
      "Accuracy at step 870: 0.9671\n",
      "Accuracy at step 880: 0.9633\n",
      "Accuracy at step 890: 0.9661\n",
      "Accuracy at step 900: 0.9657\n",
      "Accuracy at step 910: 0.9631\n",
      "Accuracy at step 920: 0.9686\n",
      "Accuracy at step 930: 0.9687\n",
      "Accuracy at step 940: 0.9682\n",
      "Accuracy at step 950: 0.9648\n",
      "Accuracy at step 960: 0.9678\n",
      "Accuracy at step 970: 0.9676\n",
      "Accuracy at step 980: 0.9688\n",
      "Accuracy at step 990: 0.9686\n"
     ]
    }
   ],
   "source": [
    "hidden1 = nn_layer(x, 784, 500, 'layer1')\n",
    "\n",
    "with tf.name_scope('dropout'):\n",
    "    tf.summary.scalar('keep_probability', keep_prob)\n",
    "    dropped = tf.nn.dropout(hidden1, keep_prob)\n",
    "\n",
    "# 创建一个输出层，输入的维度是上一层的输出:500,输出的维度是分类的类别种类：10，\n",
    "# 激活函数设置为全等映射identity.（暂且先别使用softmax,会放在之后的损失函数中一起计算）\n",
    "y = nn_layer(dropped, 500, 10, 'layer2', act=tf.identity)\n",
    "\n",
    "with tf.name_scope('loss'):\n",
    "    # 计算交叉熵损失（每个样本都会有一个损失）\n",
    "    diff = tf.nn.softmax_cross_entropy_with_logits(labels=y_, logits=y)\n",
    "    with tf.name_scope('total'):\n",
    "      # 计算所有样本交叉熵损失的均值\n",
    "      cross_entropy = tf.reduce_mean(diff)\n",
    "\n",
    "tf.summary.scalar('loss', cross_entropy)\n",
    "\n",
    "with tf.name_scope('train'):\n",
    "    train_step = tf.train.AdamOptimizer(learning_rate).minimize(cross_entropy)\n",
    "    \n",
    "with tf.name_scope('accuracy'):\n",
    "    with tf.name_scope('correct_prediction'):\n",
    "      # 分别将预测和真实的标签中取出最大值的索引，弱相同则返回1(true),不同则返回0(false)\n",
    "      correct_prediction = tf.equal(tf.argmax(y, 1), tf.argmax(y_, 1))\n",
    "    with tf.name_scope('accuracy'):\n",
    "      # 求均值即为准确率\n",
    "      accuracy = tf.reduce_mean(tf.cast(correct_prediction, tf.float32))\n",
    "tf.summary.scalar('accuracy', accuracy)\n",
    "\n",
    "# summaries合并\n",
    "merged = tf.summary.merge_all()\n",
    "\n",
    "\n",
    "def feed_dict(train):\n",
    "    \"\"\"Make a TensorFlow feed_dict: maps data onto Tensor placeholders.\"\"\"\n",
    "    if train:\n",
    "        xs, ys = mnist.train.next_batch(100)\n",
    "        k = dropout\n",
    "    else:\n",
    "        xs, ys = mnist.test.images, mnist.test.labels\n",
    "        k = 1.0\n",
    "    return {x: xs, y_: ys, keep_prob: k}\n",
    "\n",
    "with tf.Session() as sess:\n",
    "    # 写到指定的磁盘路径中\n",
    "    train_writer = tf.summary.FileWriter(log_dir + '/train', sess.graph)\n",
    "    test_writer = tf.summary.FileWriter(log_dir + '/test')\n",
    "    # 运行初始化所有变量\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    for i in range(max_step):\n",
    "        if i % 10 == 0:  # 记录测试集的summary与accuracy\n",
    "            summary, acc = sess.run([merged, accuracy], feed_dict=feed_dict(False))\n",
    "            test_writer.add_summary(summary, i)\n",
    "            print('Accuracy at step %s: %s' % (i, acc))\n",
    "        else:  # 记录训练集的summary\n",
    "#             if i % 100 == 99:  # Record execution stats\n",
    "#                 run_options = tf.RunOptions(trace_level=tf.RunOptions.FULL_TRACE)\n",
    "#                 run_metadata = tf.RunMetadata()\n",
    "#                 summary, _ = sess.run([merged, train_step],\n",
    "#                                   feed_dict=feed_dict(True),\n",
    "#                                   options=run_options,\n",
    "#                                   run_metadata=run_metadata)\n",
    "#                 train_writer.add_run_metadata(run_metadata, 'step%03d' % i)\n",
    "#                 train_writer.add_summary(summary, i)\n",
    "#                 print('Adding run metadata for', i)\n",
    "#             else:  # Record a summary\n",
    "            summary, _ = sess.run([merged, train_step], feed_dict=feed_dict(True))\n",
    "            train_writer.add_summary(summary, i)\n",
    "    train_writer.close()\n",
    "    test_writer.close()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# 补充资料  Tensorflow 可视化梯度下降/公式调参\n",
    "[Tensorflow 可视化梯度下降/公式调参](https://morvanzhou.github.io/tutorials/machine-learning/tensorflow/5-15-tf-gradient-descent/)\n",
    "来自莫烦的课程"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import tensorflow as tf\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from mpl_toolkits.mplot3d import Axes3D\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.collections.PathCollection at 0x1a87de03b70>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAAFpCAYAAABeYWb6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzt3X+wXGd93/HPF2FAaZMI8E2Dry0k\nZoyJiwOidwypOo2tQGzi1lYcSuSQFFKohjROBpd4ej1kwONMxzchDJNMnFKHcQwp9Y84RFVqZ1SC\nYOgoOPX1yAZsalDsBOuKxAogMh0rtmR/+8fdtVerc3bPnn3OOc9znvdrRqO7u+fuPWfP7tnv83y/\nz/OYuwsAAADze0HXOwAAANAXBFYAAACBEFgBAAAEQmAFAAAQCIEVAABAIARWAAAAgRBYAQAABEJg\nBQAAEAiBFQAAQCAEVgAAAIG8sKs/fOaZZ/qWLVu6+vMAAACV3X///X/n7gvTtusssNqyZYtWV1e7\n+vMAAACVmdlfV9mOVCAAAEAgBFYAAACBEFgBAAAEQmAFAAAQCIEVAABAIARWAAAAgRBYAQAABEJg\nBQAAEAiBFQAAQCAEVgAAAIEQWAEAAARCYAUAABAIgRUAAEAgL+x6BwAAQDv2HFzTh/c9oiPHjuus\nTRt17SXnaee2xa53q1cIrAAAyMCeg2u67tNf1vETz0iS1o4d13Wf/rIkEVwFRCoQAIAMfHjfI88F\nVUPHTzyjD+97pKM96icCKwAAMnDk2PGZ7kc9BFYAAGTgrE0bZ7of9RBYAQCQgWsvOU8bz9hwyn0b\nz9igay85r6M96ieK1wEAyMCwQJ1Rgc0isAIAIBM7ty0SSDVsairQzG4xsyfM7Cslj7/DzL40+Pfn\nZva68LsJAAAQvyo1VrdKunTC449J+lF3/2FJvybp5gD7BQAAkJypqUB3/4KZbZnw+J+P3LxX0tnz\n7xYAAEB6QtdYvVvSnwZ+TgAAksQSMvkJFliZ2cVaD6z+xYRtdkvaLUmbN28O9acBAIgOS8jkKcg8\nVmb2w5I+LukKd/9W2XbufrO7L7n70sLCQog/DQBAlFhCJk9z91iZ2WZJn5b0c+7+tfl3CfOg2xkA\n4sASMnmaGliZ2W2SLpJ0ppkdlvQhSWdIkrt/TNIHJb1c0u+amSSddPelpnYY5eh2BoB4nLVpo9YK\ngqgUlpChkV5flVGBV015/D2S3hNsj1DbpG5nPhAA0K5rLznvlMaulMYSMjTS58NagT1CtzMAxGPn\ntkXdeOUFWty0USZpcdNG3XjlBdEHJ9SGzYclbXok5W5nAOijFJeQoZE+H3qseoSVywEA8yprjNNI\nr4bAqiV7Dq5p+8p+bV2+W9tX9mvPwbXgfyPVbmcAQDxopM+HVGAL2iwETLHbGQAQj+F3CKMC6yGw\nagGj9QAAKaGRXh+pwBZQCAgAQB4IrFpAISAAAHkgsGoBhYAAAOSBGqsW9K0QkKUOAAAoRmDVkr4U\nArLUAQAA5UgFYiYsdQAAQDl6rDATRjgCQJ4oA6mGHivMhBGOAJCfYRnI2rHjcj1fBtLEKiKpI7DC\nTBjhCAD5oQykOlKBmEnfRjgCQBP6ljajDKQ6AivMrC8jHAGgCX0cPX3Wpo1aKwiiKAM5HalAAAAC\n6mPajDKQ6uixAgAgoDpps9hTh5SBVEdgBQBAQLOmzVJJHVIGUg2pQAAAApolbbbn4Jref+eD0aYO\n9xxc0/aV/dq6fLe2r+xneoUK6LECACCgqmmzYU/VM+6Fz9PkiLsqqcdUetJiQ2AFAEBgVdJmRUXu\no5oacVc1YJpUhE9gVY7ACgCAAk0XlE/qkWpyxF3VgIm5q+qhxgoAgDFtLOFS1iO1wUw3XnlBY71C\nVQMmljCrh8AKAIAxbcxFVVbk/pG3v67RVFvVgIm5q+ohsAIAYEwbabCd2xZ145UXaHHTRpmkxU0b\nG+2pGqoaMHW1f6mjxipCVfP6sU8oBwCpamsJly7mhpplsk/mrprd1MDKzG6R9K8kPeHury14/DWS\nfl/SGyR9wN1/M/heZqTqaA2GwQJAc6695LxTrrFSv9JgBEzNqZIKvFXSpRMe/7akX5ZEQBVA1bx+\nH9eiAoBYkAZDXVN7rNz9C2a2ZcLjT0h6wswuC7hfUWoj9VY1r88wWABoFr06qIPi9YraGHorVR+t\nwTBYAADi02pgZWa7zWzVzFaPHj3a5p+eW1upt6qjNRgGCwBAfFodFejuN0u6WZKWlpaKF0eKVFup\nt6qjNWYZ1QEAANrBdAsVtTX0Vqqe1w+R/2fKBgAAwqky3cJtki6SdKaZHZb0IUlnSJK7f8zMflDS\nqqTvk/Ssmb1P0vnu/veN7XUH+jj0likbAPQZDUd0ocqowKumPP43ks4OtkeR6mPqjZXLAfQVDUd0\nhVTgDPo29JYpGwD0FQ1HdIXAKmNt1o0BQJvqNhxJH2JezGOVMaZsANBXdeb6a2u+QvQbgVXGWLIB\nQF/VaTiyVBhCIBWYub7VjQGAVG/AEXWnCIHAqkPk8gGgObM2HKk7RQikAjtCLh8A4kLdKUKgx6oj\nDAUGgLjENl8hWY00EVh1hFw+AMQnlrpTJjhNF6nAjtQZCgwAyEOVEYp7Dq5p+8p+bV2+W9tX9lNK\nEgkCq46QywcAlJmW1aBON14EVh1hDikAQJlpWQ3m3IoXNVYdajuX30QhJMWVABDetZecd0qNlXRq\nVoM63XgRWCmP4KCJQsi2iitzOD8AMGraCEXm3IpX9oFVLiMvmpjeoY0pI3I5PwAwblJWY1qPFrqT\nfY1VLnnqJrqN2+iKzuX8AKiO0XDU6cYs+x6rXPLUZd3GLmn7yv5a6bU2uqJzOT8AqqEX+3mxzLmF\nU2XfY5XLfFJF0zsM1R2m28aUEbmcHwDV0IuN2GXfY9WHPHWV4u7RQsiiXqbjJ57R9XsfmqlIvI3l\nH/pwfgCEQy92MQb5xCP7wCq2taFmNUu3+LDbeOvy3fKC5zp2/ISOHT8x9XmKnrMpZedHWk9hpnjO\nANTHaLjTkR6Ni7kXfcU2b2lpyVdXVzv5232yfWV/4UVmcdNGHVjeMdPvFJn0PF0Zv4hI671YFG4C\n/cfn/3R1vgcwOzO7392Xpm2XfY1V6up0i0+qt5rlebrSVo0FI4+A+OQ6Gm7S9Yj0aFyyTwWmrk63\neFF67cmnT+o7T56Y6Xm60sZFhK51IF59Hw03Xi918WsW9Ef3r5Vej0iPxoUeq8TVHZm3c9uiDizv\n0GMrl+nA8g596F//02QWhW5jpCAjjwB0oWhx5U/d+42J16M2RmijOnqsIjbraL95CrlTKuJvY6Qg\nXesAulDUqCurhB5ej1K6fueAwCpSdUb7zSuV7vU2LiJ0rQPowiyNt9HrUSrX7xwQWEWqjXX4UjZ6\nERn27F1zxwPBgizmzwIwqxBzSZU16kyn9lxxPYrX1BorM7vFzJ4ws6+UPG5m9ttmdsjMvmRmbwi/\nm/khFVVNUT1CnVnkx+U68ghAPaGuRWX1Uu940+bGr0eMhA6jSo/VrZJ+R9InSx5/q6RzB//eKOm/\nDP7HHLpORaUyi2+TPXt0rQOoKtS1qKt6KUZChzM1sHL3L5jZlgmbXCHpk74+0+i9ZrbJzF7h7t8M\ntI9Z6jIVldIHjJ49AHWFbECGvBbN06ire0yUn4QTYrqFRUmPj9w+PLgPc+gyFZXSVAMs0gygjtBl\nBDFci+Y5ppCBYe4pxRCBlRXcVzg61Mx2m9mqma0ePXo0wJ/uThtvnPG5ptpqNaTUC8T8LQDqCN2A\njOFaVHZM77vjganfU6ECw6bqXlMSIrA6LOmckdtnSzpStKG73+zuS+6+tLCwEOBPd6Pvb5wYWl5V\nUWQOoI7QDcgYrkWT9n3a91SowDCljEdTQky3sFfS1WZ2u9aL1r/b9/qqLnLRbRaTpzbVAEXmAGbV\nxAChrq9FZcc0NOl7KlTRfEoZj6ZMDazM7DZJF0k608wOS/qQpDMkyd0/JukeST8h6ZCkJyX9fFM7\nG4tZ3jghAqK2i8mZxRdA33U9QKiJ62vRMY2bFOCECAy7HtEegyqjAq+a8rhL+sVge5SAqm+cUAFR\nFz1kXbe8AKBJfZzWYPSYynqumg5wUst4NIGZ12uo+sYJFRDRtQoA4XXRgGy6oTw8pvEATmonwCHj\nQWBVS9U3TqiAiK5VAOiHthrKXQY4uWc8CKxqqvLGCRUQpdC1mspM7QAwbt7r1yy/32ZDOfcApysh\npltAiVDDV2MYxjtJ36efANBf816/Zv39EN8LuU/AGTtbrz1v39LSkq+urnbyt9uUQ0/O9pX9hS2w\nxU0bdWB5x9Tfz+E1AhCnea9fdX5/nmteWe1UTI3tvjKz+919adp2pAIblkNX7Dw1AymtSwigPW01\nuOatearz+/N8L/RhTb++N6ZJBWJu88zUziy9AMa1WV4w70oTba9Ukfoo8RxKRwisSpDDrm6emoHU\nLxIAwmuzwTVvzVPbawSmtORYkRwa06QCC3SRnkq5a3SeYb1MJQFgXJsNrnmnJWh7WoMURolPkkNj\nmsCqQNs57D7UGdWtGUj9IgHkrKkGYdsNrnlrYduspU19As4cGtMEVgXajqj7UIxYV9lFQlofbZPi\nhQPIQZMNwlkaXCn39teV8qCoHBrTBFYF2o6oc+ganWT8ItGHHjyg75psEFbtleFakZ7Ue9yqILAq\n0HZEHXPXaBetwZx78IBUNN0grNIrw7UiTSn3uFVBYFWAYsR1XbUGm7pg55gyAJrSVoNw0uc2995+\nxInAqkSdiLruF3esXaN1W4PzBjBNXLBJGQBhtdEgnPa5jbm3H/kisApk3i/uGLtG67QGQwQwTVyw\nSRkAYbXRIJz2uY21tx95I7AKZJYv7lRSUnVagyECmCYu2KQMgPCabhBO+9zG2tuPvBFYBVL1i7ut\nlFSI4K1OazBUABP6gk3KAEhPlc9tjL39yBtL2gRSdZmBeafzr7LUTqi1mHZuW9SNV16gxU0bZVpf\nrX3aCuqxLrfQ9rITAOYX8nPLMmVoCz1WgVTt3ZmnR6dqb1fIeqJZW4Ox1jyQMgDSE+pzG8vglVTK\nQDAfAqtAql4A5klJVQ2YuqwnijmAIWUApCfE5zaGwSuxBHdoHoFVQFUuAPP06FQNmLquJyKAAVBV\nG704MQxeiSG4axo9cuuosWpZnbqloar1S9QTAUhBqHrQaWKo/YwhuGtSnXPZ17o3AqsO7Ny2qAPL\nO/TYymU6sLxjpvmdqgRM8wRvANCWeQfzVBVDYzOG4K5Js57LtoLqLpAKTMgs9Uuj6bhh9+w1dzyQ\ndfcsgLi01Yszb+1nV9PXpGTWc9nn1CiBVWJmrV+KrWCSHDyAoTbrQevWfoa6hsY8sCeEWc9ln1Oj\npAJ7rq2u9ir63PULYHYxpOimCXkNrVsGkoJZz2WfU6MEVj0XU6sgpiAPQPdSqAeN6Roas1nPZQpB\ndV2VUoFmdqmk35K0QdLH3X1l7PFXSrpF0oKkb0v6WXc/HHhfUUPXUy+M4gIFYFzs07PEdA2N3Szn\nss+p0amBlZltkHSTpLdIOizpPjPb6+4Pj2z2m5I+6e6fMLMdkm6U9HNN7DBmE1PBJBcooJ/6XDsZ\n0zW0b2IPquuqkgq8UNIhd3/U3Z+WdLukK8a2OV/SZwc/f67gcXQkpq72Pnf9Arnqe+1kTNdQpKFK\nKnBR0uMjtw9LeuPYNg9K+imtpwt/UtL3mtnL3f1bQfYSc4mlVdDnrl8gV7EMm2+y1yyWa2if9anX\ns0pgZQX3+djtX5H0O2b2LklfkLQm6eRpT2S2W9JuSdq8efNMO4ruhXjjz3uB6tOHD+iDGGonY5tW\nJjVdX1f7dv6qpAIPSzpn5PbZko6MbuDuR9z9SnffJukDg/u+O/5E7n6zuy+5+9LCwsIcu422xdDd\nz5IJQHxiGDbPiOP6Yri29+38VQms7pN0rpltNbMXSdolae/oBmZ2ppkNn+s6rY8QzFYfv8xjeOOz\nZAIQnxhqJ2PoNUtVDNf2vp2/qalAdz9pZldL2qf16RZucfeHzOwGSavuvlfSRZJuNDPXeirwFxvc\n56j1rUtzKIY3fqglE67f+xDpRCCQGGonGXFcXwzX9r6dv0rzWLn7PZLuGbvvgyM/3yXprrC7lqZY\nCjlDi+GNH2rJhGPHT+jY8ROS+hP4Al2qUzsZsq6HKRHqi+Ha3rfzx8zrgcUQ/Tchhu7+UEsmjEs5\nlw+kKHSanikR6mv62l6lNKZv549FmAOLIfpvQgzd/bPuQ1ErqEzqgS/QhKZGizXRs1+116zrEXCx\nafLaPktpTJ+mtCCwCqxvXZqjYnjjz7tkwpNPn9R3njxx2rapB75AaE3Wi3bVs9/XGth5NXVt72tp\nzDQEVoHF0LOD541fMMYvrFJ/Al8gpCa/FLvq2c/1i74rfS2NmYbAqgEx9OygGIEvUE2TX4pd9ezn\n+kXflb6WxkxDYDWG/Hv/Mfs7MF2TX4pdNXBy/aLvSp9LYyYhsBpB/h3T8B5BLpr+UuyiZz/XL/qu\n5JohILAaQf49rD727PAeQS76+KXYx2OKXYgAOrXvEgKrEeTfw+lrzw7vEeSkj/WifTymPkvxu4QJ\nQkfEsJhoX5T17LzvjgeSXj+R9wgAtCeGtQxnRWA1IobZxftiUg9Oyosh8x4BgPakmCUgsBrRt2n1\nuzStByf2FkcZ3iMA0J4UswTUWI0h/x5GleVkYm5xTMJ7BADakeJITgIrNGJ09E3RvDFS3C0OAED3\nUhzJae7eyR9eWlry1dXVTv422lW2jEwMKbTUhvECwKy4zoVhZve7+9K07eixQuNibXGkOIwXQLdS\nC1Kavs6l9nq0gR4rZGv7yv7CNOXipo06sLyjgz0CELOYe9/LNHmdS/H1mAc9Vi0hWk9XisN4kT6u\nGelKceWFpq5zew6u6f13PqhnxjpnYn892kBgNQdSSWmbZUFWvgzDyvX1zOma0cdznGJjrImFp4fv\n4/Ggamie16MP7xvmsZpDijPC4nlVJ/scXkTWjh2XK+0JTmOQ8+uZyzWjr+c4xTmVmpjUuOh9PKru\n69GX9w2B1RxSbL3Ebs/BNW1f2a+ty3c3vvRN1ck+c/kybEvOr2cu14y+nuMUV15oYlLjSe/XeV6P\nKu+bNr8j6iIVOIcmulhz1kWapMpkn7l8GbYl59czl2tGX89xrCOcpwk9qXHZ+3iD2VxB27T3TSqp\ndHqs5pBi6yVmsbZyU+z+j1nOr2cu14w+n+Od2xZ1YHmHHlu5TAeWd0T1hd6WsvfxR97+urlej2nv\nm1i/I8YRWM2BdePmM96lWzZDe9et3Fy+DNuS8+uZyzWjqXOcQhooB029j6e9b1LpCSUVOCfWjaun\nqEvXJBWNMem6lTtv938fRrmElGo6JZQcrhlNnONU0kC5aOJ9PO19k0oqnQlC0YmyHqrx4Cr1yeZy\nm0APaAoT+qLr62nVCUJJBaITZV23LvUqTZJKTQAQu1TSQGhOKqn0SqlAM7tU0m9J2iDp4+6+Mvb4\nZkmfkLRpsM2yu98TeF9rIxUTn7Iu3b61PvkyAMJgQl9IaaTSp/ZYmdkGSTdJequk8yVdZWbnj232\nq5LudPdtknZJ+t3QO1pXXyYc65tcCpj7PDoKaBMT+iIVVVKBF0o65O6PuvvTkm6XdMXYNi7p+wY/\nf7+kI+F2cT6kYuKUSpfuvHIJIIGmMaEvUlElFbgo6fGR24clvXFsm+sl/S8z+yVJ/0jSm4PsXQCk\nYuKVQpfuvHIfAQeExIS+SEGVwMoK7hsfSniVpFvd/SNm9iOS/sDMXuvuz57yRGa7Je2WpM2bN9fZ\n35mlMjwT/ZVDAIn09aUuiWs+ulYlFXhY0jkjt8/W6am+d0u6U5Lc/YuSXiLpzPEncveb3X3J3ZcW\nFhbq7fGMSMUAwGR9qkvimo+uVQms7pN0rpltNbMXab04fe/YNt+Q9GOSZGY/pPXA6mjIHa1rPC+/\naeMZeskZL9A1dzzAzL0AoH7VJeVSv5mDVGfan5oKdPeTZna1pH1an0rhFnd/yMxukLTq7nslvV/S\n75nZNVpPE77Lu5p5tMAwFcPMvQBwur7VJZF+T1/K39eV5rEazEl1z9h9Hxz5+WFJ28PuWniTWmWx\nnyjkpS/1LkgDdUmITcrf11nNvN63Vhn6qU/1LkgDdUmITcrf11kFVmWtL5eSyt+i3/pU74I0NF2X\nlGqtDLqT8uTKlVKBfXHtJeedtoDjUEr5W/Rbyi01pKupuqSUa2XQnaLv61R6UbPqsRptlRWhVwDz\nCtEyT7mlBoyjBxZ1pDy6M6seK+n5VtnW5btPm+VUolcA9YVqmafcUkP/zTqwgh5Y1JXq6M6seqxG\n0SuA0EK1zEO21KhtQUh1BlZwrUVusuuxGqJXAKGFbJmHaKlR24LQ6gyB51qL3GQbWLE4LkKLbS6g\nlOeBQT1Nz39Wp/HAtRa5yTawktLN3yJOsbXMqW3JSxs9lFUaD2XBHdda5CLbGisgtNhGsVDb0j+T\naubaGH03bSJRJrcFMu+xAkKLqWUeWw8a5jOtR6qNHsppaT3SzwCBFXqKtfaobembaUFLWzV+kxoP\npJ8BAiv0yDCYWjt2XCY9N09ZzqPhYupBw3ymBS0x9FDGNoAD6AI1VuiF0doOSadN/spMz0jdtJq5\nGGr8WMwZoMcKPVGUJhlHOgIpq9Ij1XUPJelngMAKPVElaCIdgZSlErR0HdwBXSOwQi+U1XYMkY5A\nH6QetDCoBDmgxgq9UFTbYYP/u55PCgBzXCEf9FihF0KmSWhVz4bXKw2znKcmzilzXCEXBFbojXnS\nJEzVUA8LPadhlvPU1DlljivkglQgssdUDfW1sYwK5jfLeWrqnLLEEnJBYIXsMVVDffRCpKHsfKwd\nO37amoNNnVPmuEIuSAUie0zVUB8zbadh0qjZ8VRfU+c0lekiqqK2EGXosUL2pn1h0KouRy9EGorO\n06jRVF+T53TntkUdWN6hx1Yu04HlHckGIoxwxCT0WCF7RTNaDwvYFxtqifaltdu3Xoi+Gj1PZT1X\nw57bLs9pKp8LRjhiEgIrZC/EF8msQ9lDjbqK4Yso9UkrczE8T9tX9k9N9XVxTlMaYUptISYhsAI0\n/1QNs3whhGrtpvRFhHhUWXOwCyn1AlFbiEmosQLmNOvw9FCt3Vn+7p6Da9q+sl9bl+8+bRQY8rJz\n26JuvPICLW7aKFP1lQmafg+l1AtEbSEmqdRjZWaXSvotSRskfdzdV8Ye/6ikiwc3v0fSD7j7ppA7\nCsRq1i+EUK3dqn+3rGdr9a+/rc/936PR17MgvFl7aNvoHU2pF4jaQkwyNbAysw2SbpL0FkmHJd1n\nZnvd/eHhNu5+zcj2vyRpWwP7CkRp1i+EeVMxw7qq8YlMy/5uWc/Wp+79BjPMtySGWrh5tJGmizVF\nWYbaQpSpkgq8UNIhd3/U3Z+WdLukKyZsf5Wk20LsXCxIo2CSWdMCdVMx0umzxI8r+rtlPVvMMN+O\nPgzNbyNNN8/nAohJlVTgoqTHR24flvTGog3N7JWStkraP/+uxYECYUxTJy1Qt7U7aZb4sqkhJk0O\nOS7GepbUpVSUXaatNB29QOiDKoGVFdxXloXYJekudy+88pvZbkm7JWnz5s2VdrBrfbgoohldpHfK\nAh+TdGB5R+Fjk+bpGhdjPUsVMafaUirKLpNamg7oUpVU4GFJ54zcPlvSkZJtd2lCGtDdb3b3JXdf\nWlhYqL6XHerDRRHhdZXeqbOQbVGK5R1v2tybUU2xp9r6sPgwaTrMK6eSmio9VvdJOtfMtkpa03rw\n9DPjG5nZeZJeKumLQfewYymNVEF7uurJrNtzUJRiWXrly6Lt5ZlF7L3KfentIU2HunIrqZkaWLn7\nSTO7WtI+rU+3cIu7P2RmN0hadfe9g02vknS7u5elCZPUl4siwuqqJ7NKPVfVtFhfviibPhfzphkZ\nmo/cxd74Ca3SPFbufo+ke8bu++DY7evD7VY8uCiiSJc9mZMCotxahlKz5yLUHGB9CWKBOnIrqWFJ\nmwq4KGJcyJ7MkIXXubUMpWZ7lZkDDJhfbiU1LGkD1BCqmDd04XVuLUOp2cJq5gAD5pfbEkD0WAE1\nhejJDNnDtOfgml5gpmcKyhz72jIcaqpXmTnAgPnlVlJDYAV0KFQP07Dnqyio6nPLsGk5zAEGtCGn\nkhpSgUCHQs1xVDYj+wYz5huaQ9/nAAMQHj1WQIdCFV6X9XA9605QNac+zwEGIDwCK6BDo7UHa8eO\na4PZKYXQVb+scxt107U+pTViXg4ISBGpQKBjO7ctPjdqZlgjNevowNxG3SCM2JcDAlJEYAVEYNLo\nwCpYy+10Oa1NVte87zsApyMVCEQgxOjA2NNTbaaccpyBvo4c5z0DmkZgBUSg7zVSbQU6w+Ct6LXs\n+wz0dfT9fQd0gVQgEIGyGqmLX7PQi3RWGymn0XqhMvTEnIraPCA8eqyACBTNTHzxaxb0R/ev9SKd\n1UbKqWwur1H0xJwqtxmxgTYQWAGRGK+R2r6yvzcLKreRcpoWpNETUyz22jwgNaQCgUj1qbC4jZTT\npCCNUZIA2kKPFRCpGAuL647sayPlVDaLPQEVgDYRWAEtmiUwCbXcTSjzjuxrOuVEvRCAGBBYAQ0Z\nD6JmLUaPLVCYNLKviSkT6hwz9UIAukZgBTSgqHfnU/d+Qz623bTAJKZAYZaar7rBERN7Akhdr4vX\nWdICXSnq3RkPqoZSKUYvq+0av79o/blr7nhAWyp8DlliBUDqettjRcsXXZolWGq7GL1ub1LVmq9J\nQeW0z2HMIyHbXJIHQLp622NFyxddKguWbOx228XoRb1J1336y5V6c6su9DwtCJr0OazaK9a2eV43\nAHnpbWAVc8sX/Vc2b9M73rR5amDSpHkbHDu3LerA8g49tnKZDizvKNz3KkFQ2ecw1iVWaKgBqKq3\nqcAY5wBCPmIb0TfURoOjKGU4ruxzmPPr1hRSmEC7ehtYxTYHEPIT04i+oTYaHKPB0dqx4zKdWrg/\n7XOY6+vWBGpNgfb1NhVYtR4EyElbqbZhyvCvVi7TR3/69cl/DmNNUU5DChNon7mXDQJv1tLSkq+u\nrnbyt4GcVU0NkUI6VROvR1MkL91UAAAQPElEQVSv8fB5i3rZpPVBFI+tXDb33wG60sX1yczud/el\nqdsRWAEYN55Cklh3L7SmXuOi5x23uGmjDizvqP03gC51dX2qGlhVSgWa2aVm9oiZHTKz5ZJt3m5m\nD5vZQ2b232fdYQDxIIXUvKZe46LnHZVCChOYJPbr09TidTPbIOkmSW+RdFjSfWa2190fHtnmXEnX\nSdru7t8xsx9oaocBNG/aKLiU0mKxKnuN144d156Da7WPfdJIxcUMXlf0X+yjdKuMCrxQ0iF3f1SS\nzOx2SVdIenhkm38v6SZ3/44kufsToXcUQHsmjYJrYqRZl6PXugroyl5jSXMde9nzkv5DX8Q+SrdK\nKnBR0uMjtw8P7hv1akmvNrMDZnavmV0aagcBtG/SKLgmuuG76trvckb1otd46PiJZ/S+Ox6otcZp\nqiMYgapif49X6bEaX4VDOn092RdKOlfSRZLOlvS/zey17n7slCcy2y1ptyRt3rx55p0F0I5JE3Ve\nc8cDhb9zZJDCqtP701XXfllAd/3ehxrvxRo+3/tKXk+pXs9drJOsAqHE/h6fOirQzH5E0vXufsng\n9nWS5O43jmzzMUn3uvutg9uflbTs7veVPS+jAoE0bV/ZX9gNv2njGXrq5LMzjdSZNi1A0+mrrct3\nn9ZKLNLkiKOy13PUBjN95O2vi+aLA8hRyFGB90k618y2mtmLJO2StHdsmz2SLh784TO1nhp8dLZd\nBpCCsm54M82UzhtNwxVpo2u/ak1Gk2nJSSnBoWfcWfQZSMTUwMrdT0q6WtI+SV+VdKe7P2RmN5jZ\n5YPN9kn6lpk9LOlzkq519281tdMAulO2qsGxJ08Ubl+Wzps0LcCkGdr3HFzT9pX92rp8d+UapLLf\nqRLUTDuOeY2+npPENJwcQDkmCAUQRFlKqyydV5aGmzQreJ2JAaf9znhd2JNPn9R3CoLENkbVVZ3c\nM6Z6EiAXVVOBvV2EGUC7Zl34vM6Q6UmjB8sCjbLfef+dD0o6fdHnskCsjRFHw/14/50P6pmSRi8L\nKQNx6+0izADaNevC53WGTNcZPVj2WFndUtcLuO/ctqiPvP11E1OUpAWBeNFjBSCY8d6fadtKsw2Z\nLuvleoFZ6WzlkybiLOvtqnocTU0uOvralO17LLNMAzgVgRWAzswSiEnF6Ubp+d6n4XNW+Z2hugFK\n07PFD1+bstq1WGaZBnAqUoEAkjFM022w0+ctLkuPTfodqX6AMmly0VlHLU4S+yzTAE5FYAUgKTu3\nLerZksLust6nsrqleQKUsr917PiJoEvkTKv5qjP9BIDmkAoE8JyuFiSeVZ0RhaGXwZhUuzVq2qjF\nKspSpl0uXg2gGIEVAElxfkmXBXqzTu0wNGtN1yTTardGNVVoXmf6CQDNIrACICm+L+kqgV6XvWtF\n+1A2uWhTheZdLV4NoByBFQBJ8X1JTwv0QvY+1dX15KJ1UqIAmkXxOgBJ5V/GXX1JxxboVdH25KKM\nGATiQ48VAEmzL0nTtFR7Y9rsSYshJQrgVARWACSVf0lL6wsst/3FHVugF6sYUqIAnkdgBeA502qG\n2hwpSG8MgBQRWAEo1fVIQXpjAKSG4nUApVIsIAeALtFjBaBUqgXksUhlJnsA4dBjBaAUw/nrG9an\nhVw3EED8CKwAlGp7XqY+mVSfBqC/SAUCmIgC8nqoTwPyRI8VADQgtpnsAbSDwAoAGkB9GpAnUoEA\n0AAmOAXyRGAFAA2hPg3ID4EVkAjmRAKA+BFYAQnocs0+AEB1FK8DCWBOJABIQ6XAyswuNbNHzOyQ\nmS0XPP4uMztqZg8M/r0n/K4C+WJOJABIw9RUoJltkHSTpLdIOizpPjPb6+4Pj216h7tf3cA+Atlj\nzT4ASEOVHqsLJR1y90fd/WlJt0u6otndAjCKOZEAIA1VAqtFSY+P3D48uG/cT5nZl8zsLjM7J8je\nAZDEmn0AkIoqowKt4D4fu/0nkm5z96fM7L2SPiFpx2lPZLZb0m5J2rx584y7CuSNOZEAIH5VeqwO\nSxrtgTpb0pHRDdz9W+7+1ODm70n6Z0VP5O43u/uSuy8tLCzU2V8AAIBoVemxuk/SuWa2VdKapF2S\nfmZ0AzN7hbt/c3DzcklfDbqXANACJmEFMK+pgZW7nzSzqyXtk7RB0i3u/pCZ3SBp1d33SvplM7tc\n0klJ35b0rgb3GQCCYxJWACGY+3i5VDuWlpZ8dXW1k78NAOO2r+wvnNJicdNGHVg+rWQUQGbM7H53\nX5q2HTOvA4CYhBVAGARWAKDyyVaZhBXALAisAERvz8E1bV/Zr63Ld2v7yn7tObgW/G8wCSuAEKqM\nCgSAzrRVVD58rrqjAhlRCEAisAKguIOCD+975Lmgauj4iWf04X2PBN/HupOwMqIQwBCBFZC52IOC\nporKQwaTbQZ/AOJGjRWQuUlBQQyaKCofBpNrx47L9XwwWbd2ixGFAIYIrIDMxR4UNFFUHjqYZEQh\ngCECKyBzsQcFO7ct6sYrL9Dipo0yrU/YeeOVF8yVYgsdTDKiEMAQNVZA5q695LxTaqyk+IKCukXl\nZc7atLFwlvW6weS8IwoB9AeBFZC5HIOCJoLJ0MEfgDQRWAHILijIMZgE0A4CKwBZyi2YBNAOitcB\nAAACIbACAAAIhMAKAAAgEAIrAACAQCheB4AExbxwNpAzAisASEzsC2cDOSMVCACJiX3hbCBnBFYA\nkJjYF84GckZgBQCJiX3hbCBnBFYAkJhrLzlPG8/YcMp9sS2cDeSK4nUASAxrHQLxIrACgASx1iEQ\nJ1KBAAAAgRBYAQAABEJgBQAAEEilwMrMLjWzR8zskJktT9jubWbmZrYUbhcBAADSMDWwMrMNkm6S\n9FZJ50u6yszOL9jueyX9sqS/CL2TAAAAKajSY3WhpEPu/qi7Py3pdklXFGz3a5J+Q9I/BNw/AACA\nZFQJrBYlPT5y+/DgvueY2TZJ57j7/wy4bwAAAEmpElhZwX3+3INmL5D0UUnvn/pEZrvNbNXMVo8e\nPVp9LwEAABJQJbA6LOmckdtnSzoycvt7Jb1W0ufN7K8kvUnS3qICdne/2d2X3H1pYWGh/l4DAABE\nqEpgdZ+kc81sq5m9SNIuSXuHD7r7d939THff4u5bJN0r6XJ3X21kjwEAACI1NbBy95OSrpa0T9JX\nJd3p7g+Z2Q1mdnnTOwgAAJCKSmsFuvs9ku4Zu++DJdteNP9uAQAApIeZ1wEAAAIxd5++VRN/2Oyo\npL9u6c+dKenvWvpbseHY88Sx54ljzxPH3o5XuvvUkXedBVZtMrNVd89ymR2OnWPPDcfOseeGY4/r\n2EkFAgAABEJgBQAAEEgugdXNXe9Ahzj2PHHseeLY88SxRySLGisAAIA25NJjBQAA0LjeBFZm9m/M\n7CEze7ZoncKR7S41s0fM7JCZLY/cv9XM/sLMvm5mdwyW70mCmb3MzD4z2PfPmNlLC7a52MweGPn3\nD2a2c/DYrWb22Mhjr2//KOqpcuyD7Z4ZOb69I/f3/by/3sy+OPhsfMnMfnrkseTOe9nnd+TxFw/O\n46HBed0y8th1g/sfMbNL2tzvECoc+380s4cH5/mzZvbKkccK3/+pqHDs7zKzoyPH+J6Rx945+Ix8\n3cze2e6ez6/CsX905Li/ZmbHRh5L/bzfYmZPmNlXSh43M/vtwWvzJTN7w8hj3Z13d+/FP0k/JOk8\nSZ+XtFSyzQZJfynpVZJeJOlBSecPHrtT0q7Bzx+T9AtdH9MMx/4bkpYHPy9L+vUp279M0rclfc/g\n9q2S3tb1cTR57JL+X8n9vT7vkl4t6dzBz2dJ+qakTSme90mf35Ft/oOkjw1+3iXpjsHP5w+2f7Gk\nrYPn2dD1MQU+9otHPtO/MDz2we3C938K/yoe+7sk/U7B775M0qOD/186+PmlXR9TyGMf2/6XJN3S\nh/M+2P9/KekNkr5S8vhPSPpTSSbpTZL+Iobz3pseK3f/qrs/MmWzCyUdcvdH3f1pSbdLusLMTNIO\nSXcNtvuEpJ3N7W1wV2h9n6Vq+/42SX/q7k82ulftmPXYn5PDeXf3r7n71wc/H5H0hKSpE9xFqvDz\nO7bN6Gtyl6QfG5znKyTd7u5Puftjkg4Nni8VU4/d3T838pm+V9LZLe9jU6qc9zKXSPqMu3/b3b8j\n6TOSLm1oP5sw67FfJem2VvasBe7+Ba13ApS5QtInfd29kjaZ2SvU8XnvTWBV0aKkx0duHx7c93JJ\nx3x9wenR+1PxT9z9m5I0+P8Hpmy/S6d/+P7zoCv1o2b24iZ2siFVj/0lZrZqZvcOU6DK7Lyb2YVa\nb/X+5cjdKZ33ss9v4TaD8/pdrZ/nKr8bs1n3/91ab8kPFb3/U1H12H9q8F6+y8zOmfF3Y1V5/wep\n362S9o/cnfJ5r6Ls9en0vFdahDkWZvZnkn6w4KEPuPv/qPIUBff5hPujMenYZ3yeV0i6QNK+kbuv\nk/Q3Wv/SvVnSf5J0Q709DS/QsW929yNm9ipJ+83sy5L+vmC7Pp/3P5D0Tnd/dnB31Oe9QJXPabKf\n8Skq77+Z/aykJUk/OnL3ae9/d//Lot+PUJVj/xNJt7n7U2b2Xq33Wu6o+Lsxm2X/d0m6y92fGbkv\n5fNeRZSf96QCK3d/85xPcVjSOSO3z5Z0ROvrDG0ysxcOWrnD+6Mx6djN7G/N7BXu/s3BF+gTE57q\n7ZL+2N1PjDz3Nwc/PmVmvy/pV4LsdCAhjn2QBpO7P2pmn5e0TdIfKYPzbmbfJ+luSb866C4fPnfU\n571A2ee3aJvDZvZCSd+v9VRCld+NWaX9N7M3az3o/lF3f2p4f8n7P5Uv2KnH7u7fGrn5e5J+feR3\nLxr73c8H38PmzPK+3SXpF0fvSPy8V1H2+nR63nNLBd4n6VxbHwn2Iq2/Eff6erXb57ReeyRJ75RU\npQcsFnu1vs/S9H0/LQc/+FIe1hztlFQ4AiNSU4/dzF46THOZ2ZmStkt6OIfzPnif/7HW6xD+cOyx\n1M574ed3bJvR1+RtkvYPzvNeSbtsfdTgVknnSvo/Le13CFOP3cy2Sfqvki539ydG7i98/7e25/Or\ncuyvGLl5uaSvDn7eJ+nHB6/BSyX9uE7trY9dlfe8zOw8rRdpf3HkvtTPexV7Jf3bwejAN0n67qDB\n2O15b6tKvul/kn5S61HqU5L+VtK+wf1nSbpnZLufkPQ1rUftHxi5/1Vav9AekvSHkl7c9THNcOwv\nl/RZSV8f/P+ywf1Lkj4+st0WSWuSXjD2+/slfVnrX6z/TdI/7vqYQh67pH8+OL4HB/+/O5fzLuln\nJZ2Q9MDIv9enet6LPr9aT19ePvj5JYPzeGhwXl818rsfGPzeI5Le2vWxNHDsfza49g3P897B/aXv\n/1T+VTj2GyU9NDjGz0l6zcjv/rvB++GQpJ/v+lhCH/vg9vWSVsZ+rw/n/Tatj2Q+ofXv93dLeq+k\n9w4eN0k3DV6bL2tkRoAuzzszrwMAAASSWyoQAACgMQRWAAAAgRBYAQAABEJgBQAAEAiBFQAAQCAE\nVgAAAIEQWAEAAARCYAUAABDI/wf0Xo3AZVKeAAAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 720x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "LR = 0.1\n",
    "\n",
    "# 真实参数\n",
    "REAL_PARAMS = [1.2, 2.5]\n",
    "# 测试参数\n",
    "INIT_PARAMS = [[5,4],\n",
    "              [5,1],\n",
    "                [2,4.5]][2]\n",
    "\n",
    "# 生成x数据\n",
    "x = np.linspace(-1, 1, 200, dtype=np.float32)\n",
    "\n",
    "# test 1  简单的线性函数\n",
    "# y_fun = lambda a, b: a * x + b\n",
    "# tf_y_fun = lambda a, b: a * x + b\n",
    "\n",
    "# test2  稍微复杂一点的函数,经验公式调参\n",
    "# y_fun = lambda a, b: a * x**3 + b * x**2\n",
    "# tf_y_fun = lambda a, b: a * x**3 + b * x**2\n",
    "\n",
    "# test3 解释局部最优\n",
    "y_fun = lambda a, b: np.sin(b*np.cos(a*x))\n",
    "tf_y_fun = lambda a, b: tf.sin(b*tf.cos(a*x))\n",
    "\n",
    "noise = np.random.randn(200)/10\n",
    "# 生成y数据，（训练标签数据）\n",
    "y = y_fun(*REAL_PARAMS) + noise\n",
    "\n",
    "#展示数据\n",
    "plt.figure(figsize=(10,6))\n",
    "plt.scatter(x,y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "# 初始化数据，这里的a，b想当于网络中的w，b，初始化之后，随着网络的训练会对其进行改变\n",
    "a, b = [tf.Variable(initial_value=p, dtype=tf.float32) for p in INIT_PARAMS]\n",
    "pred = tf_y_fun(a,b)\n",
    "mse = tf.reduce_mean(tf.square(y-pred))\n",
    "train_op = tf.train.GradientDescentOptimizer(LR).minimize(mse)\n",
    "\n",
    "a_list,b_list, cost_list = [],[],[]\n",
    "with tf.Session() as sess:\n",
    "    sess.run(tf.global_variables_initializer())\n",
    "    for t in range(400):\n",
    "        # 记录a_,b_的变化记录，以及mse的值的大小\n",
    "        a_, b_, mse_ = sess.run([a, b, mse])\n",
    "        a_list.append(a_)\n",
    "        b_list.append(b_)\n",
    "        cost_list.append(mse_)\n",
    "        \n",
    "        result, _ = sess.run([pred, train_op])\n",
    "        \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "a= 1.2104619  b= 2.5165994\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[<mpl_toolkits.mplot3d.art3d.Line3D at 0x1a87df428d0>]"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXcAAAD8CAYAAACMwORRAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz\nAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBo\ndHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJztnXmYFNW5xt8zC8uwCAwiIMwMqDGo\nUSJj4hZXbiQkbolGEJdEEwJucY0a1HiNRK96o1GDShQXhqjElRiNqLgkKsoYcQEussigoLIjwz7T\n7/3j9FLTU9VV1bV0dfX3e55+Zrrq1KmvTp96z1ffWUqRhCAIghAvygptgCAIguA/Iu6CIAgxRMRd\nEAQhhoi4C4IgxBARd0EQhBgi4i4IghBDRNwFQRBiiIi7IAhCDBFxFwRBiCEVhTpx7969WVdXV6jT\nC4IgFCXvvffeGpK72qUrmLjX1dWhsbGxUKcXBEEoSpRSTU7SSVhGEAQhhoi4C4IgxBARd0EQhBgi\n4i4IghBDRNwFQRBiiIi7IAhCDBFxFwRBiCEi7oIgCDFExF0QBCGGiLgLglAyTJsG1NUBZWX677Rp\nhbYoOAq2/IAgCEKYTJsGjB0LbNmivzc16e8AMGZM4ewKCvHcBUEoCSZMyAh7ii1b9PY4IuIuCEJJ\nsHy5u+3Fjoi7IAglQU2Nu+3Fjq24K6WmKKVWKaU+ttg/Rin1YfLzllLqAP/NFARB8MbEiUBVVdtt\nVVV6exxx4rk/BGBEjv2fAjiS5P4Afg9gsg92CYIg+MqYMcDkyUBtLaCU/jt5cjw7UwEH4k7yDQDr\ncux/i+T65NfZAAb4ZJsgCCVMEMMWx4wBli0DEgn9N67CDvg/FPJcAC/4nKcgCCVGqQ1bDALfOlSV\nUkdDi/uVOdKMVUo1KqUaV69e7depC04pTYwQhDAIe9hiHO9hXzx3pdT+AO4H8AOSa63SkZyMZEy+\nvr6efpy70IiHIQj+E+awxbjew549d6VUDYCnAJxJ8hPvJhUXpTYxQhDCIMxhi3G9h50MhXwUwNsA\n9lZKfa6UOlcpNU4pNS6Z5DoA1QAmKaXmKqUaA7Q3cpTaxAhBCIMwhy3G9R52MlpmNMl+JCtJDiD5\nAMl7Sd6b3P8Lkj1JDk1+6oM323/yjbmV2sQIQQiDMIctxvUelhmqyMTcmpoAMhNzcyLwpTYxQhDC\nIqxhi3G9h0Xc4S3mVmoTIwQhbsT1HlZkYQat1NfXs7ExGuH5sjLtsWejlPYaBEEQooJS6j0n4W/x\n3FHYmFscx9cKglB4RNxRuJibl1i/IAhCLkTcUbiYW1zH1wpCMROXp2mJuRcQifULQrTInq0K6Kf4\nKHWwSsy9CIjr+FpB8EIhPec4PU2LuBeQuI6vFYR8KXQ/VJxmq4q4F5C4jq8VhHxx4jkH6dnH6Wla\nxL3AlNLLAwTBjlye87RpQO/ewBlnePfsrRqIkSO1o2WkWJ+mpUNVEITIUFenBTub6mpg69b2Xn2K\n2lrtHDnBqtP07LOBhx9uu10pYNw4YNIkp1cQPNKhKghC5LALqVj1QwHWwg64i4lbhX4mT26/nQSe\nf9553lFCxF0QhFBw0llq1Q+1zvItzho3MXGrhqC11V36qCNhGUEQQsEq5OIkpGJ1LOB+HLpVXuXl\n5gLvJuQTBhKW8Qmrx8i4zGIThLDwMszQLFwD6Fi82xFmVqGfsWNjNjSZZEE+w4YNY9RpaCCrqkj9\nEKk/VVXk+PHm2xsaCm2xIESX2tq290zqU1vr7PiGBp1WKf3Xy/1mlZef5wgKAI10oLElEZaZNk13\noixfrmNzEyc6a+mL/fFNEKJEMUztLwYkLJPEy4y3Uul4EYQwkEl74RJ7cfeyVoRVD3x5ubv0Ep8X\nBI1M2guP2Iu73504bjteCr1WhiCEhTgx0SL24u5lrQirx8hJk5w/XsZplTlBsEKcmOgR+w7VQnfi\nyJrtQingZAx7vgMbhLZIh2qSQnfixGmVOUGwwi78KZ59+MRe3IHCduLImu1CKWDnxEh4MnxKQtzd\n4mfHUKGfHAQhDOycmDi9BKNYEHHPIojHRxn+JcQdOycmqPCkjNCxJvYdqm7xsriRIAjmBDGwIde6\n7M8/H9+OW6cdqiLuWcjoFkEIBr9Hy1g5Ykq1vYfjtsSBjJbJE78eH90+LsrjpRB3/A5PWsXrs52z\nUu24jZW4+yGQfoxucRu39xrnl4ZBKEX8eEFHrHGydGQQH7+X/LVanjefJTu9LvvpdmlTL0uh+nnd\nguCEqCyLa1b3lfK2rHAxAIdL/sZG3L2uFe0nVhUMcJdeKftzRem6hfgTNWciu6EphXctOBX32IRl\nCjWO1iwkkutxUan2oRMvcX4ZPyyESdQmI2XH8VPrPlVXZ9J07lwY2wpNbMS9ENP8rWLlI0dqEbci\nO6buJc5vdX2kxN8F/ykWZ2Lr1sz/a9eW6FIHTtz7ID5Rjrk7JVdIxCosYxU6yTeOaXbd+ZZBVGKp\nQnSJahjQWHfLy6Npo1+g1GLuZPjilCtW7kTgncTUnZC6bieNSK484h6rFLwTlXpivNerq8kOHcK7\n3wpNSYp7kJg1HLm8GDuPOghPQjpmhTAo9BOek3srznVZxN1HrLwVu555YwOQLbxBeDteBNpLwyAI\nfuGk4XAa9iz000VQ+CbuAKYAWAXgY4v9CsCdABYD+BDAgU5OXEzibuehO/FiwvB2rDya6mr784nn\nLhQapyGfXEONjZ/ycmf3W6GfRNzip7gfAeDAHOI+EsALSZE/GMA7Tk5cTOLup1cbdEVqaNBi7tZz\niUosVYgXbuq7UwfDieeeqrt25y/Geu9rWAZAXQ5xvw/AaMP3hQD62eUZtrh7EVXHXm1LC7liBfnJ\nJ+SiReTq1WQi0caGMCpSvl54sXkwQrRxW9+dOlFm+VZWkoN6rucgLOXR/f+PT93RxL8+tN32/G7u\nlajcH2GK+3MADjd8fwVAvV2eYYq7V1G1Ov7RKVvIZ58lf/lLcr/9yIqK9rWkSxfyoIPIiy/mmX3+\nyUpsDzz0IfFzIQoEuQxHwyOtPLHvbN6Aa/lWx6O4rWuvdge2QnExBnMaRvNMPMyeWJsOU6bw0qAU\nysMPU9z/YSLuwyzSjgXQCKCxpqYmjHIg6V/MPBXu6IcVvLvTZdzWpWf7TPv0IffYgxw8mNxll3b7\n16Inb8ZvOBBNgYmuxM+FKODWyXAkoOvXk7fdRtbVtc+4qkpX8r32Ivv3ZwvK2uzfio58GGdyP3yY\nztNrKKgQ95SEZQzk6oBx2ho3NJC9OzfzJlzJLeiUEevab5M33ki++Sa5eXP7A9euJV9+mbzmGs6v\n/Fb6uJ0o56M4jXtgke8VJEpehlC65COIVs7W9HvW8P5uF3MTumQyGjiQvPBC8plndDjUEAIlyT1q\ndnB/zOX5uIsv4Vi2QgtBKxSf6jKGXLHCcyduIZ6GwxT3H2Z1qL7rJM8wxd2qkrmZyTZqt1lchpp0\nouk4hcMwx5UwNzSQ3+v4DhtwOndAh3C2oQM//tGV5Ndf+3S1mXNFIT4olC6+OBk7d/Lds+/mWmSe\nkmdiOE/u+BwbHmm1Pb/x3HVYyj/hQm5DcsZT9+7k/fezYWoi7+GXRe25A3gUwBcAdgL4HMC5AMYB\nGJfcrwD8GcASAB85ibczZHG3qmRW3nyb1ri1lfzd79Kt/nv4Ng/CO3m33CnRHYDPOL3LzzIn7deP\nr1zxggiyECs8ORkffUQecED6HnkJx/IAvO9KWM1GjtVgGWd2Pj6zYdQosrnZ9jqi8jQsk5iycDvD\nlKT+wX/8YxJgC8p4Pa5jOXb623K/8w753e+mM/wjLmZHbJVQihAbnAh8mzQ1Cb571l1kJx3+XIo6\nnownCSRyhkTMzmMpylMT5COPkF276o1Dh5Kffeb5OsJAxN0BOVvjDRvIww7TG3fZhS9fOTO4lrul\nhTf3uCkdqpmL/bkHFpk2HlGpYELp4KXOOfF4jWm6YwOfhcGrPvdcDhm4yTYkkus8Oe1fsIDcc8+k\nS1+jhzCHVDb5IuLuENMfZ8MGsr5eF8+AAeT8+dZpfUIpsh7v8hPoirYGvXgEXmvjnUTp0VAoDbzW\nOSex6lSawVjMeRhCAlyHHhzX+2+ObfAUE1+7ljz0UH1A377kwoWhlE2+lLy45y3Ezc3k4YeTAL/u\nM5iH7r4slFY5VTm7YSOfw0gS4HZU8vLqKe3SRKFTRygNvNY5J6NMlCK/h9e5GjpA/hH2ZR2WtnNs\nct3PnkezbNpEHnOMPmjgQHLZMttDCnU/lrS4592itrSQJ5xAAmzuNYB7d1oWWqtstLkMLfxfXJI5\n8Q03kIlEziGdEqIR7MjH4cklml4W+TIK4Nhdn+R2VJIAn8NIdsNG1yLpi9A2N2c8+CFD9Jj6HBRq\neGRJi3veP/Tll+uEPXrwmP7zQ2+Vs2+W2efcR5YlJ2JccQVraxKmNkmIRrAjX4fH6l6qrnaWn+15\nH3mErWV6TPKduIBlaMmrLvsWIlm/ntx3X53B8OHkzp2uy0Y89wDFPa8W9YkndKKKCnLWrEBbZVce\n1PTp6WUNFh47jl06t+YUeFk/RjDDy3pDZqJpNsTQKj/LOnfPPekDPzj5OtbWJDzVS9/q9qef6pnm\nAPnb3+Y8n8TcQxZ3u4qcXQme/uPSzFIBf/qTozzyxenoAaN9sy7/R3pY2CfH/CqnB5+r8ZEO2dLF\ni7NiXHoj5bXnU//aYBB23nqrp2sLhNde00/NSpEvvmiZTEbLhCzudsOijPsqsZ3vln1HfznppPQU\n5qCE0EnDY3bel347Ky3wvOQSS4HP1fhIh2zp4uW3N6uTVo2Fo7r08MOZA+6+29uFBckNN2gb+/Qh\nV64stDVpSlrcSesWNbuS34rLSICfldeS69a1Oa66Wn+ctspOWnE7DyrnTfjCC3ptU4AfnniN68Yn\nSutjCOHixVmxqpN5vV1s+vRMP9Jtt/lxacHR0pIZQXPMMfp7BCh5cbfCWCEPw79IgDtQwYPxtuMb\nwNVMuKxj7TwoWwF+6qn0ojiNp9/m6pFQPPfSJt8Qgt0orez8LM/z4ouZZbGvv97nqwuIlSsz8fc7\n7yy0NSRF3C1JCVwHbEtPmPhvXMvaWmfi57WDya4RcCTAxhWRHn3U8bVLzF3IBzdOgVUd+8eN/8lM\n9b/ssnYrOFoRiQEAzzyj7e7alVy+vAAGtEXE3YJU5bsW/00CXIC92bPzVjY0OAtbWFV0q49ZyCNX\nhXUswLfckmylOpCvvurq+gt+swiRJzs82aGDM6fA7P6oxaf8qryv/jJ6tF6Mz6ENUXFGmg7Sa0w9\ni+NZW5PIOQBCRssUcPmBGbcsSC/7+dPdXnPlNTt9Oa+XkIejypJI6LWsAT3SZ9489ycSBBPMRLWy\n0ln/U/b90QPrOB/f1F+OPprcts2xHVEJIzY0kHt0+pwb0J0E+GM8YTlAI4wGSMTdikSCPOIIfem/\n+EWbXV7WsDCb1KEUOX68d5Mtxb6lJb1qJQcPJtes8X4yoeTxIqrGY8uxky/iv0iACyr3azPj04/B\nB2GRuqZxmEQCXIF+7I4NjkO5fiPibsXjj+vL7tOHXLeu3W67SperARg/Ps8RBDmwbXA2byYPPJBp\nz2jHjvxPJgj0PiY+VV//iItJgKuwK5++Y5lpmlz3SVQ891R5KLTyTRxCArwJV1KpwjRAIu5mbNum\nPVyAvO++vLNxOszSj8roKM/ly8nddtM7Lrgg/5MJAr3X44YG8vLqB0joxe9evPZfeeUflZi70d6D\n8A4J/T7WQ3ZvEs/d7FMQcb/jDn3JQ4bkXDMiX4JoxR3n+dZbmV4vDw2XEG+chEM8i+q//52ej8G/\n/KXdbjf3SRQGAGSXx6M4jQS45PCzJOZu9gld3NevJ3v10pc8Y0YgpyiY557ioYf0zooK8vXX8z+p\nEEvcCFHeotrUlBkXfuGFpkmiEm5xg7E8Du+/hC3llfrL++/LaJnsT+ji/pvf6Ms98kjHY2zd4rYV\n99uLamggJ3e7lAS4pqw3n/pT4cfkCtEhCFE11uFv1GzlmkHDdKbHHmv5dByVcIsnUb4kuST38OFt\nliwJQ+RF3I18/jnZsaO+3Dlz0puD+DGc5um3F5XKzzhCYXbZwfzrQ9sDu1ahuPA7bJhdhydhHAlw\n0651tiO3Cl0fPTcwa9eSPXroA2fODLXBEnE38utf60s95ZT0pkJ7D357Ucb8qrGayzGABDil20U5\nF34SoS8dgqxzp0PPmt6GDvxh30YfrQ4GX8riD3/QB33ve6GGmkTcU3z5Jdm5s77UuXPTmwsd9/Pb\ni8rO77t4O/12m/N7P2Z6rkI+Egvh48ShceNRp+rcEMxjM3TGY3FvUSxE53W4Z20t2R0bub6sJwnw\nCLzu6/2cCxH3FKlY+wkntNlc6AkSQXpRqc8FuJME+DW6cm8syCnwUe7MEtxjJdK+LH2RpLaW7IJN\n6TWapmIMgUQ63yiHAfO9/7LL6DpcTwJ8tXK4eO6pTyjivmZNZrGid99ts6vQnrvfYSHT/Don+OnB\no0iAH2MfdsEmS3EvBm9LcEa+dcvtPdEwNcHHy0e3qV9VVXoyXxQ6THPhVxn1wDpuRDcS4JEd3w7l\nmkXcSfLaa/UlHndcu12FjrmnbPDTuzHNb9MmbuivPauHcaZ47iVAvo6L66fZSXo6frPqwiGY38Zj\nL4Y6ls/9Z1ZGE3E1CfDzoT+U0TIMUNyN8bCNKvnqvH//O2faqD46+sa8edzZUbdkZ+JhR0sklEzZ\nxJB8Q46uRHnOnMykuWnTfDl/MWBWRr2xiptV0lP8z38Ct6Ekxd3ojV8EPRv1X2XfK0ph8l1cH9DT\nwdmlC5+9dWHe6+cI0ceveLLl775hAzlokE5w3nm+nb8YsCqj+SP0Ojo866w2aYNwkEpS3FOVqgwt\nXAy9hsyJeLroKlUg4ppI6LW0AXLo0JxLr8b55iwFvNQfW0Ey1qMDDzStR3F3DkzLaOlS/frAykry\niy8CLYOSFPfU4+DxeJYEuASDWIaWonsctBLX8nKPlWPjxszCaRddZJkszo/VpUJgYbXUEhddupAL\nF4Z//ihz8sm6bH73u0AdpJIU91SBvoKjSYAX449F6XHmeiGI59Z/zpzMok7PPmuaRDx3wZSFC7Wo\nA+SDDxbamujx2mu6bPr0YSdsDcxBciruZYgREycC3+n0IY7Bq9iErpiCc1BVpbcXEzU11vu2bAEm\nTPCQeX09cPPN+v+f/xz47LN2SSZOBKqq2m4rxnIUfGTHDmD0aGDzZmDUKODsswttUfQ44ghg6FBg\n1SqcX/2YaZJc97bvOGkBgvgENVpm0ZHnkADvxIVF+zhoFq/ztfVvbSVHjiQBvtPxcFZgp+mElpJ7\nrBasuewyXfnq6nSHqg+EXcfyOZ/rYx58kAS4tnYoqzonJObuG6tX6wXClCI/+cQ0SbGIVkODjrEH\nFR7526RVXKH6kwD/G9fGrtNL8JF//pPpTp+33vIly7A7XfM5X15vVtu6Nb3k8cwJr8loGd+47TZ9\nSSNHmu4utl78IO2trSWPxKtshWILyngEXks3HMXSANoR1euIil2O7Pjyy8xbvm680bdzh92v43r2\nbYN135etjddcoxOefrq/F5Gk9MQ9kSD33ltf0jPPmCYpxo7CfNYIcUKq4v4eE0iAyzGAvbAm3YAU\nSwNoRVQb8qjY5ciO1lZyxAi986ij9AvZfSLsEVluz2elFXY2NjSQh+6+jK1Q3IaOnH7vWt8b89IT\n9zfe0JfTr5/lSwLiMsTPD4FIVd4K7Ei/9PcpnMTysoRpGUW5ATQjqg15VOxyZMcf/6g39upFfvZZ\n+Ocv4PlyjVjL5e2n7ssXcBwJ8JKyO9ITef1qzEtP3M84Q1/Ob39rmSQqN5ZX/LgOY0WsxafcgO4k\nwF/hnlg0gFFtyKNil60d772XGTJreBL2ywuNeszd6h5TytkxJ+NJEnpBNaC9w+RFc0pL3NetIzt1\n0pezZIllsqg8EudLrkWZ8hEI4416fu/HSIBbVSfui49sK2NU4sZWRLUhj4pdOe3YtIn8xjf0BsPy\nAkGsZBrV0TJm16qU7mS1wthgVmAHv4DuqzgEb/ramJeWuN+p1y3n8OG2SaMuSlbYDY/0RSDO0cNI\n56l92QlbLG/gYmgko2pjVOzKaUeyHnC//cgtW9LHRKVhCgu3WpFdPjfhShLgFPwsup47gBEAFgJY\nDOAqk/01AF4F8D6ADwGMtMvTq7inCx4Jzq/8lr6U6dM95RllcnnsvglEc3O6U/rhrudZVupiucmj\n2pBHxS5TOx7TT3Ds1In86KM26XOFcqJyTYUku8HcA4tIgJvRmd2xwbd71TdxB1AOYAmAwQA6APgA\nwD5ZaSYDGJ/8fx8Ay+zy9SLuxkKsx7skwFXYNf0y6Dhi18Hj2830n/9klnJ9+mlXthQ6ni145NNP\nye6674WTJrXbbdWoV1dH42kkCmQ3cl/scwwJ8Ope90ZvtAyAQwC8aPh+NYCrs9LcB+BKQ/q37PL1\nIu7GSvYnXEgCvB2/jpzn6Cehesu3364ztxglUSyeu+CCHTvIgw/WP+RJJ+mhxVlYhXKqq6U+WDJ1\nqi6Mww7zLUs/xf0UAPcbvp8J4O6sNP0AfATgcwDrAQyzy9eLuKc8xwrs4FfYlQR4IBpj7TmGGqtN\nJNLLE/DII9uNb45K3Fhoi6fQyAQ934G7765fT+niHPIkl4NNmzI3S47BHm7wU9xPNRH3u7LSXArg\nMmY89/kAykzyGgugEUBjTU1N3heX8hxH4jkS4DwMIZDwviRuxAk1rvnVV5mZib//fWFtEWzx1ODO\nmqV/SKX0yoYukSe59hjvjye7JIdp33CDL3mHHZaZB2Cg4ftSAH1y5etHzP1RnEYCvAp/EA8yCGbO\n1IVaXk6++abjw0T4wydvgV29muyv1xjitdfmdW55kmtLdnn8F14kAW7su5dpuMstfop7RVKsBxk6\nVPfNSvMCgJ8l/x8CYCUAlStfr6NlHp+8gVugx7bXYJl4DEHxm99kCnX9etvkcqMXhrxCI4kEefzx\nOuGhh1rO7HaCNOgZshvacuzkSvTVX2bP9py/30MhRwL4JDlqZkJy2w0ATkj+vw+AN5PCPxfA9+3y\n9DzOfcoUEuCrONK0Ugs+sX07WV+vC/XUU209D3lELwx5lftdd5EAt1ftwkN3XybC7BNmDe1tuFT/\nc8EFnvOP/ySmo/Xbls7FX0y9FamgPrJoEdm1Kwnwil735xQB6VwrDK6fmObO1ctjAxzTYbo8afmI\nWUM7FP/R//TurUcmeSDe4r58OakUWyo7sgfWi6cYAv8ep4d0NaOKe2OBpQjks7SqPM77g9MVRB+9\nv5n85jdJgNO6/lLuH58xbWg7J7h+wL76y4wZnvKPt7jPmKFL79RTTSumeIr+U1tLPgLd6/8+DmDH\n5DsizdaccepBSnw+eMzK+IHyX+p/hgxhFTbL/RMApg3tzTfrwQl/+IOnvOMt7qQeP7p8ucR4Q0Ip\nshs2chH2IKEnjVmJgFNvXH475+T7hJNdxqdgOglwKzryuH4fWDpH8hsEwNq1+uUnHom/uCcR7y8c\nUiJRj3e5AxUkwJF4zpMISHzeGqOYV1cz7zXBjWVch6Vcj11IgOfhbkthl/sn2pSMuJMSt7XDj/Ix\nNqJX4H9IgKvQm0/ctTJvu8RzN8fJCqBOyylVxh2wjXMwjAT4NE6k2RrjqTzl/ok2JSXugjkNDebr\nfuTrmaUaiTK08o1Ow3VmxxyT9+vX5KnLHLsVQN084aTK+C6cTwJcijr2wDp5YipiRNxLHDvvz7N3\nvHJl+i3vnDDBk53y1NWWXCuA5vMb/usCvYzvNnTgj/rOkYW+ihwR9xLHzvvzxUubNYssK9MZPvec\nDxkKpDPP3fETzsKF6TkKvPtukvLEVOw4FfcyCLFk+fLc+2tqfDjJ0UcDN96o/z/zTGDZMh8yFSZO\nBKqq2m6rrASqqwGlgNpaYPJkYMwYm4y2bgVOOQVobgZ++lPgvPMA6OMmT9b5GPMDgLo6oKxM/502\nze8rE0LFSQsQxEc892DJ5f0ZvTTje1nLyzOP5469uNZW8kc/0gcOG0Zu2xbMBZkQ55COk2uzm7R0\nP84lkVywauNG2/OJN18cQMIypY1VzL26uq0IWMXlXd3Y69aRgwbpA8eNC+yajJS6GFld//jx+u9Z\neIgEuAWd+N1Oc12//1Pi8NFFxF2w9f7sYruubuz33kuvVcKpU327BitKXYysrr+sjNwfc7kZnUmA\n5+B+R+VSyDkHcX4CCwIRd8EWu1EZrm/s++7TB3bu3O7lyn5T6hOgrK6/F9ZwKepIgA/ibKbGs9uV\nS6Eay1J/AssHp+IuHaoljF2nqtNO12nTkh1xv/olnuxylu7I+/GPgfXrPdvo1jZfOoqLALPrLEcL\nHsMoDMIyzEE9xuMeAMoyvRGrTtzm5mA7WCdMALZsabttyxa9XfCIkxYgiI947oXHacw912Nzdh6d\nsZkfqAP0l+OOa/MCCD8fv0vd42toaP+b3YLLSYBfog8HYLnrcvFryQM3lPoTWD5AwjKCE+xGy9iJ\nqNnjfA2WcXWZfnE5L73UNp98Rb/UY7XGyUijMY0EuAMV/FH319uIdHW19wXHggrTlHrfST6IuAu+\nYHfzWXle38MbZIVeYIwPP2w5K7K6ur3oV1bmJ0hRx+/GKNVgHoD30x2ol1Te5bhhzkVYHnWpP4Hl\ng4g7xbPzA7ubPKf433svCbClogO/i7dN0zn5xOFmD0rEnrxrBVeUDyABPt7l52yYmnkNohevOEyP\nWu5Td5S8uItH4A9265DYlvP48STAlejL3fFZ3gIfxmN6kCITiFg2N5MHHqgzOvRQcuvWNru9eN9y\n/0SXkhd3ieV5p6FBh0iyy7BDh/adqpaiuGMHZ+EoEuBc7M9u2NguLONE3IPuYAtazHwPc7S0kCee\nqDMZPJhctapdEq/3gHjU1hSybEpe3KUX3jtW4lBd7S6fAwas4QLsTQKcieGswA4CmTXHnKyCGHSj\nnK8QFuytU5deqjPo0YNcsMDSNvG+/afQ5Vry4i6eu3fcNJB2wyX36bSEX0IvEWycXGPMM9VwhDEE\nz8u1Gq/LaWewr4Jwzz06g4pANbygAAASq0lEQVQKvTJnDsT79p9Ca0vJi3uhW9c44LQSOynrhgby\n+L7vshk64fW4zjLfQghSPjes26V5fbmuGTMyY1YffDCPDASvFDoqUPLiTorX4hWnDaQrYfz739kC\nHY9JrXsShZBZPs6A3y/VsOXVVzPr91xzjU+ZCm4Rzz0C4i54x0kD6TZ8Mw46rNCCMv4Ef4tMyMyt\nM+Dn6/BsaWwku3XTGY4bRyYS9scIgVDoqICIuxAabjyZVNrrcR0JPaNyJJ6jUsX3ZJVr+QZfPboF\nC8jevXVmo0aZvrNWnlLDRUbLiLiXBE7Wjk+R8fITvBWXkQC3oiOPxUuFMd4jga/H0tREDhyoM/vB\nD8jt201tkP6l0kHEXQiVhgbzMevZItPWy0/wz9CTnDarKvJf//LNFq9eVSTWu/nsM3LPPXVhHXYY\nuXmzabJCx4CFcBFxF0InVww6JXTZXqZCKx8p/5n+0q0bOXu2Jxu8erFOG6nAaWrSk5MAPQt1/XrL\npIUevSGEi1Nxl/XcBd/I9VLupiZg7Fj9v/HlzDW1ZSh/8H7gtNOATZuA4cOBN97I2wYv64NPm6Zt\nXLu2/b5Q1xhvagKOOgpYuhQYNgx4+WWgRw/L5KW+tr1ggZMWIIiPeO7xw8noEctQwY4d5OjROlHn\nzuQ//5mXDV68WDv7Q/GE/+//yJoafcKDDsrpsaeQmHtpAfHchbAxe5tPNpbefWUlMHUqcO65+k1O\nJ5wAPPlku2Tptz5ZvB3Iixeb68nDaR6eaGwEDj9cG3LIIcDMmTk99hRjxrR9Gqqt1d/HjAnYXiHa\nOGkBgviI5178mHUeGl/+kVcnX2sr+etfZ1zl229vcz4nM2Hz9WJz2R24J/zSS2TXrkyPirHoPBUE\nSIeqEARG8c4OgWRPtc87VJBIkDfdlDnw4ovJ1lZXyyGYjVixG8niZkinr9x3X+bFJmPG6BCVIFgQ\na3GXCRuFwcmkHaPQev6djGsOH388d8EGy1i4sdHJ53WBvtnshpYW3XClDLriCv3kIgg5iK24S+dR\n4XDSYeql09FUWGfNInv2JAEurvgGv4n5pp613Yu+gx4L7rpRWLOGHDFCG1FZST7wgD+GCLEntuIu\nEzYKR5DrrudstJcsIb/1LRLg1+jKU/F4mzR2L/xIiW5QI2BcOxxvvZWZdVpdTb7+uncjhJIhtuIu\nEzYKh53n7uUJyrbRbm4mTzstvWMKfs4hAzexocG+0Ul500E5BY7zbmkhb701E18/+GBy+XLX55Ow\nZGkTW3EXz71wmHmoKWH1KjKOGu1Egpw0iezUSe/cc0/yrbdsG52UbUGF8xzZvmgRefjh6Z3zR1zC\nPWu2uxbo8eNzd2QL0cTPBjm24i4x98ISlNeYq9HOPuffb/44HaahUlw4/Dz27Wze2er7yzJc2s6d\nO8k77shU2r59+eqlM/Kqw7meUsS5iS5+a5av4g5gBICFABYDuMoizU8BzAcwD8Bf7fKU0TKCEasb\nYPx48+1/nbKVvOqqdIhjc8/+vKi6gQqtlqNlgqozVra/dPUr5H77ZTaefjq5dq1lY1BentuuXE8o\nEpaMLn5HG3wTdwDlAJYAGAygA4APAOyTlWYvAO8D6Jn83scuXxnnLmRjJsBWN0Z1td73LXzI9zoc\nnNnx7W+TM2e2y9cvz8nJGPoR/eZy+bCTMicbNIh89tl0Hrn6CHLZles48dyji9/9hH6K+yEAXjR8\nvxrA1VlpbgHwCycnTH1E3AUnOBmho9DKcR0e4Oaeu2c2HnYY+dRTZEuLb56TXZ/D879vJE8yiHpV\nFXnjjeTWrW3ysesjqK42P7/VccX4opNSIsqe+ykA7jd8PxPA3VlpnkkK/JsAZgMYYZHXWACNABpr\namryuzKhpHD6KjuA7IQtvKnHzdxW1SO9ceNue/Iy3Mr++Nyz52RmSxWaeQ7u57uoNxjSSS+hsGKF\naT5OJoOZzYq1alzGj8+raIWQiGzMHcCpJuJ+V1aa5wA8DaASwCAAnwPokStf8dwFJzh9lZ3x06fz\n15xzxh3ctGtdemMrFF/CsTwfdyUnQiVce04pL31XfMWz8SCfwkncjM7pc6xDD97X/TLyiy8cXVeq\nb8BNiEb6m4qTQoyWUTqtNUqpQwBcT/K45PerkwuO3WRIcy+A2SQfSn5/JdnxOscq3/r6ejY2NuY8\ntyAAeuXHCRP0Yok1NUBzs/ma60Zqa4FytmD/5X/HGWjAj/AcOmJHev9K1R+tB34HGwfsi8lv7ov3\n1wxEh/7VOG9CL/zktApg+3Y88/h2PHzLV+j4ZRMO6NGEvbe8j6E73sFgfNrmXG/iUNyHX+FvOBXb\nVWckEs6v64wzcqeprgbWrHGWn1AaKKXeI1lvm86BuFcA+ATAsQBWAJgD4HSS8wxpRgAYTfJspVRv\n6M7VoSQtb0ERdyFfUi/VyH4phxGl9N9U9e6B9TgRz+L7mIlj8Qp2w6q8z78ZVfg3DsczOAkzcAJW\nYvf0vtpaYNky53n17m3fUDU0yPK9QgbfxD2Z2UgAd0CPnJlCcqJS6gbox4MZSikF4H+hh0y2AphI\n8rFceYq4C15IefNNTeb7a2v1X7P9tTXEshcW4KIjP0CfNfOwD+ajL75EL6xDNdaiXCWws6wjtrVW\nYjV2xXLUoAm1WIAheAffxXzsixZUQKlM4wHotewnT9b/G580Jk60FmcnDZXbBkOIN07F3TZuE9RH\nYu6CH+TqrLLryHIyEifX0EOr9ezdzuK1em9rvh2/QrxBXGeoCkI2uTqrcu3LNRLHrrPTSnC9rL9j\nJfCpMf3SiSqQIu5CTAh6ZqmdEOfy3M3wsnKmmddfWUl26OC8gRDij1Nxl3eoCpElFY9uatKy1tSk\nv2e/NzWffOvqgDPP1O9iNSP1HtLq6vb7qqp0HN0ML+9qNXsXavfuwI4dbdNt2aJj+oKQEyctQBAf\n8dwFO4JYAdTJuHmz97I6fXpw+7YqO2SJayEbiOcuFDtWHq7VdidMmGA+MqW8POMtT57cdnTLmDF6\ntEoiof/mGpZo9L6BzJDMFLm8fjOsngScPCEIpY2IuxBZrASM1GGVfMIzVg1DIuFMvJ2QagxIYOrU\ntmGW7IbDjokTdYNgxG0DIZQmIu5CZDETthT5xt/D9oTdeP1Wx2fH4d02EEJpIuIuRJbsEEc2+XQs\nFqMn7LWBEEoTEXch0qSELTt2ncJt/L1QnnBqhE5ZWf4hJUFwQ0WhDRAEJ9TUmC8lkE84ZcyYcL3f\n7CUGUiGllC2CEATiuQtFQTGGU1KYjdCRsepC0Ii4C0VBMXcsBjGkUxDskLCMUDSEHU7xCz9DSoLg\nFPHchUCRjsTiDikJxYuIuxAYQa0NU2wUc0hJKF4cvawjCORlHfGnrs7iZRm18vIJQcgXpy/rEM9d\nCAzpSBSEwiHiLgSGLHolCIVDxF0IDOlIFITCIeIuBEZcOxJlBJBQDMg4dyFQinVsuhWylIBQLIjn\nLggukKUEhGJBxF0oSfINrcgIIKFYEHEXIkmQcW0vk6vcjACS2LxQSETchcgR9MzWfEIrKaFuanL2\nXlSZnSsUGpmhKkSOoGe2lpVpwc1GKf22o2yyO1FTaUlt08SJ7TtTZXauEBROZ6jKaBkhcgQd13a7\nSqOZp58Sdiuhlti8UGgkLCNEjqBntrqdXJWPUMvsXKHQiLgLkSPoma1uJ1flI9QyO1coNCLuQuQI\nY2Zr6sXbiYT+myvvfIQ6rrNzheJBOlQFwQHTpunY+/Ll2mM360QVhDCQDlVB8JG4LaMgxB8JywhC\nRJFJUIIXxHMXhAgiC5QJXhHPXRAiiCxQJnhFxF0QIohMghK8IuIuCBFEJkEJXhFxF4QIIpOgBK+I\nuAtCBJFJUIJXHIm7UmqEUmqhUmqxUuqqHOlOUUpRKWU7wF4QhNy4mUUrCNnYirtSqhzAnwH8AMA+\nAEYrpfYxSdcNwEUA3vHbSEEQBMEdTjz37wBYTHIpyR0AHgNwokm63wO4BcA2H+0TBEEQ8sCJuO8O\n4DPD98+T29Iopb4NYCDJ53y0TRAEQcgTJ+KuTLalVxtTSpUBuB3AZbYZKTVWKdWolGpcvXq1cysF\nQRAEVzgR988BDDR8HwBgpeF7NwD7AXhNKbUMwMEAZph1qpKcTLKeZP2uu+6av9WCIAhCTpyI+xwA\neymlBimlOgAYBWBGaifJjSR7k6wjWQdgNoATSMp6voIgCAXCVtxJtgC4AMCLABYAmE5ynlLqBqXU\nCUEbKAiCILjH0aqQJJ8H8HzWtuss0h7l3SxBEATBCwV7E5NSajUAk3fQu6I3gDU+mOMnUbQJELvc\nEkW7omgTIHa5wQ+baknadloWTNz9QCnV6OR1U2ESRZsAscstUbQrijYBYpcbwrRJ1pYRBEGIISLu\ngiAIMaTYxX1yoQ0wIYo2AWKXW6JoVxRtAsQuN4RmU1HH3AVBEARzit1zFwRBEEyIvLgrpU5VSs1T\nSiVyrRNvteZ8cmbtO0qpRUqpx5OzbL3a1Esp9VIyz5eUUj1N0hytlJpr+GxTSp2U3PeQUupTw76h\nXm1yalcyXavh3DMM230vK6d2KaWGKqXeTv7WHyqlTjPs86287N5NoJTqmLz2xcmyqDPsuzq5faFS\n6rh8bcjTrkuVUvOTZfOKUqrWsM/09wzJrp8ppVYbzv8Lw76zk7/5IqXU2SHadLvBnk+UUhsM+wIp\nK6XUFKXUKqXUxxb7lVLqzqTNHyqlDjTsC6ScQDLSHwBDAOwN4DUA9RZpygEsATAYQAcAHwDYJ7lv\nOoBRyf/vBTDeB5tuAXBV8v+rAPyPTfpeANYBqEp+fwjAKQGUlSO7ADRbbPe9rJzaBeAbAPZK/t8f\nwBcAevhZXrnqiSHNeQDuTf4/CsDjyf/3SabvCGBQMp9yn8rHiV1HG+rP+JRduX7PkOz6GYC7TY7t\nBWBp8m/P5P89w7ApK/2FAKaEUFZHADgQwMcW+0cCeAF6IcaDAbwTZDmRjL7nTnIByYU2yUzXnFdK\nKQDHAHgime5hACf5YNaJybyc5nkKgBdIbvHh3Llwa1eaAMvKkV0kPyG5KPn/SgCrAPi9upyTdxMY\nbX0CwLHJsjkRwGMkt5P8FMDiZH6h2EXyVUP9mQ29gF/QOH2XgxnHAXiJ5DqS6wG8BGBEAWwaDeBR\nH86bE5JvQDtwVpwI4BFqZgPooZTqh+DKKfri7hCrNeerAWygXh/HuN0ru5H8AgCSf/vYpB+F9hVs\nYvLx7HalVEcfbHJjVyell16enQoVIbiycmMXAEAp9R1or2yJYbMf5WX7bgJjmmRZbIQuGyfH5ovb\nvM+F9gJTmP2eYdr1k+Rv84RSKrWCbFDl5TjfZOhqEIBZhs1BlZUdVnYHVq8crS0TNEqplwH0Ndk1\ngeSzTrIw2cYc2z3Z5OR4Qz79AHwLeuG1FFcD+BJawCYDuBLADSHaVUNypVJqMIBZSqmPAHxtks7x\nUCqfy2sqgLNJJpKb8y6v7OxNtmVfo+91yQGO81ZKnQGgHsCRhs3tfk+SS8yOD8CuvwN4lOR2pdQ4\n6KeeYxweG5RNKUYBeIJkq2FbUGVlR+j1KhLiTnK4xyys1pxfA/34U5H0wrLXos/LJqXUV0qpfiS/\nSIrRqhxZ/RTA0yR3GvL+IvnvdqXUgwAud2KTX3Ylwx4guVQp9RqAbwN4EnmWlV92KaW6A/gHgGuS\nj66pvPMuryzs3k1gTPO5UqoCwC7Qj9tOjs0XR3krpYZDN5ZHktye2m7xe/ohWLZ2kVxr+PoXAP9j\nOPaorGNfC8MmA6MAnG/cEGBZ2WFld1DlFJuwjOma89Q9Fq9Cx7wB4GwATp4E7JiRzMtJnu1ifkmB\nS8W5TwJg2sMehF1KqZ6psIZSqjeAwwDMD7CsnNrVAcDT0HHJv2Xt86u8cr6bwMTWUwDMSpbNDACj\nlB5NMwjAXgDezdMO13Yp/SrL+6DflbDKsN309wzRrn6GrydALwsO6CfV7yft6wng+2j79BqYTUm7\n9obuoHzbsC3IsrJjBoCzkqNmDgawMem0BFVORTFa5mTo1m07gK8AvJjc3h/A84Z0IwF8At0KTzBs\nHwx9Ey4G8DcAHX2wqRrAKwAWJf/2Sm6vB3C/IV0dgBUAyrKOnwXgI2iRagDQ1aeysrULwKHJc3+Q\n/HtukGXlwq4zAOwEMNfwGep3eZnVE+gQzwnJ/zslr31xsiwGG46dkDxuIYAf+FzP7ex6OVn/U2Uz\nw+73DMmumwDMS57/VQDfNBx7TrIcFwP4eVg2Jb9fD+DmrOMCKytoB+6LZB3+HLpfZByAccn9CsCf\nkzZ/BMPIv6DKSWaoCoIgxJC4hGUEQRAEAyLugiAIMUTEXRAEIYaIuAuCIMQQEXdBEIQYIuIuCIIQ\nQ0TcBUEQYoiIuyAIQgz5f7Ks/gnxOOmaAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAcUAAAE1CAYAAACWU/udAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDkh3KscAAAAASUVORK5CYII=\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# 可视化代码\n",
    "# 这一步是看拟合结果，即我们获得的a，b能否拟合数据\n",
    "print('a=',a_,' b=',b_)\n",
    "plt.figure(1)\n",
    "plt.scatter(x,y, c='b')\n",
    "plt.plot(x, result,'r-', lw=2)\n",
    "\n",
    "# 3d cost figure\n",
    "# 3d视角看误差如何下降\n",
    "fig = plt.figure(2)\n",
    "ax = Axes3D(fig)\n",
    "# 两个参数轴形成的面\n",
    "a3D, b3D = np.meshgrid(np.linspace(-2,7,30),np.linspace(-2,7,30))\n",
    "# 计算对于每个参数值对，，mse的值，即mse在（a,b)的空间分布\n",
    "cost3D = np.array([np.mean(np.square(y_fun(a_,b_) -y)) for a_,b_ in zip(a3D.flatten(), b3D.flatten())]).reshape(a3D.shape)\n",
    "ax.plot_surface(a3D, b3D, cost3D, rstride=1, cstride=1, cmap=plt.get_cmap('rainbow'), alpha=0.5)\n",
    "# 画出初始点\n",
    "ax.scatter(a_list[0], b_list[0], zs=cost_list[0],s=300, c='r')\n",
    "ax.set_xlabel('a');ax.set_ylabel('b')\n",
    "# 画出轨迹\n",
    "ax.plot(a_list, b_list, zs=cost_list, zdir='z', c='r')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "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": "3.6.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}