怎样用条形码查阅雅诗兰黛真假_百度知道
怎样用条形码查阅雅诗兰黛真假
我有更好的答案
国内雅诗兰黛专柜很多，也容易鉴别真假，您到专柜看看产品实物亲，条形码是最容易仿制造假的，回来再对比你的东西，真假不是看条形码，而是产品本身、包装等
采纳率：61%
所以系统录入的信息不全。如果扫不出来，一般是因为该商品用的人少，任何人都能复制。　　明确一点：条形码并不能用作防伪。　　二维码一般也不能做为防伪依据，有一种情况可以：如果同一款产品的每一个包装的二维码信息都不同（有加密编号！不要被不良商家忽悠了。因为一般条形码的信息是公开的，冒充真品。条形码扫描出的结果实际上是商品编号，APP作用就是查询这个编号对应的商品信息而已，所以很多不良商家复制别家商品条形码贴在自己的山寨商品上，制作成本高），即商品编号是唯一的（这些编号都被商家存入数据库，并且有查询记录）。　　所以买东西最好是到正规渠道　　你可以使用我查查、微信等APP扫描，一般是可以查询到商品价格、信息的
本回答被网友采纳
看包装，扫条码，闻闻味道!和真的比较比较!
商浮沉此延早操纬
其他1条回答
为您推荐：
其他类似问题
雅诗兰黛的相关知识
换一换
回答问题，赢新手礼包
个人、企业类
违法有害信息,请在下方选择后提交
色情、暴力
我们会通过消息、邮箱等方式尽快将举报结果通知您。利用微信监管你的TF训练a year ago这段代码里面，我所做的修改主要是：0.导入了itchat和threading1. 把原本的脚本里网络构成和训练的部分甩到了一个函数nn_train里def nn_train(wechat_name, param):
global lock, running
with lock:
running = True
# mnist data reading
mnist = input_data.read_data_sets("data/", one_hot=True)
# Parameters
# learning_rate = 0.001
# training_iters = 200000
# batch_size = 128
# display_step = 10
learning_rate, training_iters, batch_size, display_step = param
# Network Parameters
n_input = 784 # MNIST data input (img shape: 28*28)
n_classes = 10 # MNIST total classes (0-9 digits)
dropout = 0.75 # Dropout, probability to keep units
# tf Graph input
x = tf.placeholder(tf.float32, [None, n_input])
y = tf.placeholder(tf.float32, [None, n_classes])
keep_prob = tf.placeholder(tf.float32) #dropout (keep probability)
# Create some wrappers for simplicity
def conv2d(x, W, b, strides=1):
# Conv2D wrapper, with bias and relu activation
x = tf.nn.conv2d(x, W, strides=[1, strides, strides, 1], padding='SAME')
x = tf.nn.bias_add(x, b)
return tf.nn.relu(x)
def maxpool2d(x, k=2):
# MaxPool2D wrapper
return tf.nn.max_pool(x, ksize=[1, k, k, 1], strides=[1, k, k, 1],
padding='SAME')
# Create model
def conv_net(x, weights, biases, dropout):
# Reshape input picture
x = tf.reshape(x, shape=[-1, 28, 28, 1])
# Convolution Layer
conv1 = conv2d(x, weights['wc1'], biases['bc1'])
# Max Pooling (down-sampling)
conv1 = maxpool2d(conv1, k=2)
# Convolution Layer
conv2 = conv2d(conv1, weights['wc2'], biases['bc2'])
# Max Pooling (down-sampling)
conv2 = maxpool2d(conv2, k=2)
# Fully connected layer
# Reshape conv2 output to fit fully connected layer input
fc1 = tf.reshape(conv2, [-1, weights['wd1'].get_shape().as_list()[0]])
fc1 = tf.add(tf.matmul(fc1, weights['wd1']), biases['bd1'])
fc1 = tf.nn.relu(fc1)
# Apply Dropout
fc1 = tf.nn.dropout(fc1, dropout)
# Output, class prediction
out = tf.add(tf.matmul(fc1, weights['out']), biases['out'])
return out
# Store layers weight & bias
weights = {
# 5x5 conv, 1 input, 32 outputs
'wc1': tf.Variable(tf.random_normal([5, 5, 1, 32])),
# 5x5 conv, 32 inputs, 64 outputs
'wc2': tf.Variable(tf.random_normal([5, 5, 32, 64])),
# fully connected, 7*7*64 inputs, 1024 outputs
'wd1': tf.Variable(tf.random_normal([7*7*64, 1024])),
# 1024 inputs, 10 outputs (class prediction)
'out': tf.Variable(tf.random_normal([1024, n_classes]))
biases = {
'bc1': tf.Variable(tf.random_normal([32])),
'bc2': tf.Variable(tf.random_normal([64])),
'bd1': tf.Variable(tf.random_normal([1024])),
'out': tf.Variable(tf.random_normal([n_classes]))
# Construct model
pred = conv_net(x, weights, biases, keep_prob)
# Define loss and optimizer
cost = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=pred, labels=y))
optimizer = tf.train.AdamOptimizer(learning_rate=learning_rate).minimize(cost)
# Evaluate model
correct_pred = tf.equal(tf.argmax(pred, 1), tf.argmax(y, 1))
accuracy = tf.reduce_mean(tf.cast(correct_pred, tf.float32))
# Initializing the variables
init = tf.global_variables_initializer()
# Launch the graph
with tf.Session() as sess:
sess.run(init)
# Keep training until reach max iterations
print('Wait for lock')
with lock:
run_state = running
print('Start')
while step * batch_size & training_iters and run_state:
batch_x, batch_y = mnist.train.next_batch(batch_size)
# Run optimization op (backprop)
sess.run(optimizer, feed_dict={x: batch_x, y: batch_y,
keep_prob: dropout})
if step % display_step == 0:
# Calculate batch loss and accuracy
loss, acc = sess.run([cost, accuracy], feed_dict={x: batch_x,
y: batch_y,
keep_prob: 1.})
print("Iter " + str(step*batch_size) + ", Minibatch Loss= " + \
"{:.6f}".format(loss) + ", Training Accuracy= " + \
"{:.5f}".format(acc))
itchat.send("Iter " + str(step*batch_size) + ", Minibatch Loss= " + \
"{:.6f}".format(loss) + ", Training Accuracy= " + \
"{:.5f}".format(acc), wechat_name)
with lock:
run_state = running
print("Optimization Finished!")
itchat.send("Optimization Finished!", wechat_name)
# Calculate accuracy for 256 mnist test images
print("Testing Accuracy:", \
sess.run(accuracy, feed_dict={x: mnist.test.images[:256],
y: mnist.test.labels[:256],
keep_prob: 1.}))
itchat.send("Testing Accuracy: %s" %
sess.run(accuracy, feed_dict={x: mnist.test.images[:256],
y: mnist.test.labels[:256],
keep_prob: 1.}), wechat_name)
with lock:
running = False
这里大部分是跟原本的代码一样的，不过首先所有print的地方都加了个itchat.send来输出日志，此外加了个带锁的状态量running用来做运行开关。此外，部分参数是通过函数参数传入的。然后呢，写了个itchat的handler@itchat.msg_register([itchat.content.TEXT])
def chat_trigger(msg):
global lock, running, learning_rate, training_iters, batch_size, display_step
if msg['Text'] == u'开始':
print('Starting')
with lock:
run_state = running
if not run_state:
threading.Thread(target=nn_train, args=(msg['FromUserName'], (learning_rate, training_iters, batch_size, display_step))).start()
msg.reply('Running')
作用是，如果收到微信消息，内容为『开始』，那就跑训练的函数（当然，为了防止阻塞，放在了另一个线程里）最后再在脚本主流程里使用itchat登录微信并且启动itchat的服务，这样就实现了基本的控制。if __name__ == '__main__':
itchat.auto_login(hotReload=True)
itchat.run()
但是我们不满足于此，我还希望可以对流程进行一些控制，对参数进行一些修改，于是乎：@itchat.msg_register([itchat.content.TEXT])
def chat_trigger(msg):
global lock, running, learning_rate, training_iters, batch_size, display_step
if msg['Text'] == u'开始':
print('Starting')
with lock:
run_state = running
if not run_state:
threading.Thread(target=nn_train, args=(msg['FromUserName'], (learning_rate, training_iters, batch_size, display_step))).start()
msg.reply('Running')
elif msg['Text'] == u'停止':
print('Stopping')
with lock:
running = False
elif msg['Text'] == u'参数':
itchat.send('lr=%f, ti=%d, bs=%d, ds=%d'%(learning_rate, training_iters, batch_size, display_step),msg['FromUserName'])
param = msg['Text'].split()
key, value = param
print(key, value)
if key == 'lr':
learning_rate = float(value)
elif key == 'ti':
training_iters = int(value)
elif key == 'bs':
batch_size = int(value)
elif key == 'ds':
display_step = int(value)
通过这个，我们可以在epoch中途停止（因为nn_train里通过检查running标志来确定是否需要停下来），也可以在训练开始前调整learning_rate等几个参数。实在是很简单……812收藏分享举报{&debug&:false,&apiRoot&:&&,&paySDK&:&https:\u002F\u002Fpay.zhihu.com\u002Fapi\u002Fjs&,&wechatConfigAPI&:&\u002Fapi\u002Fwechat\u002Fjssdkconfig&,&name&:&production&,&instance&:&column&,&tokens&:{&X-XSRF-TOKEN&:null,&X-UDID&:null,&Authorization&:&oauth c3cef7c66aa9e6a1e3160e20&}}{&database&:{&Post&:{&&:{&isPending&:false,&contributes&:[],&title&:&利用微信监管你的TF训练&,&author&:&coldwings&,&content&:&之前回答问题\u003Ca href=\&https:\u002F\u002Fwww.zhihu.com\u002Fquestion\u002F2Fanswer\u002F?group_id=882752\& class=\&internal\&\u003E在机器学习模型的训练期间，大概几十分钟到几小时不等，大家都会在等实验的时候做什么？ - Coldwings 的回答 - 知乎\u003C\u002Fa\u003E的时候，说到可以用微信来管着训练，完全不用守着。没想到这么受欢迎……\u003Cp\u003E这里折腾一个例子。以TensorFlow的example中，利用CNN处理MNIST的程序为例，我们做一点点小小的修改。\u003C\u002Fp\u003E\u003Cp\u003E首先这里放上写完的代码：\u003C\u002Fp\u003E\u003Cdiv class=\&highlight\&\u003E\u003Cpre\u003E\u003Ccode class=\&language-python\&\u003E\u003Cspan\u003E\u003C\u002Fspan\u003E\u003Cspan class=\&ch\&\u003E#!\u002Fusr\u002Fbin\u002Fenv python\u003C\u002Fspan\u003E\n\u003Cspan class=\&c1\&\u003E# coding: utf-8\u003C\u002Fspan\u003E\n\n\u003Cspan class=\&sd\&\u003E'''\u003C\u002Fspan\u003E\n\u003Cspan class=\&sd\&\u003EA Convolutional Network implementation example using TensorFlow library.\u003C\u002Fspan\u003E\n\u003Cspan class=\&sd\&\u003EThis example is using the MNIST database of handwritten digits\u003C\u002Fspan\u003E\n\u003Cspan class=\&sd\&\u003E(http:\u002F\u002Fyann.lecun.com\u002Fexdb\u002Fmnist\u002F)\u003C\u002Fspan\u003E\n\u003Cspan class=\&sd\&\u003EAuthor: Aymeric Damien\u003C\u002Fspan\u003E\n\u003Cspan class=\&sd\&\u003EProject: https:\u002F\u002Fgithub.com\u002Faymericdamien\u002FTensorFlow-Examples\u002F\u003C\u002Fspan\u003E\n\n\n\u003Cspan class=\&sd\&\u003EAdd a itchat controller with multi thread\u003C\u002Fspan\u003E\n\u003Cspan class=\&sd\&\u003E'''\u003C\u002Fspan\u003E\n\n\u003Cspan class=\&kn\&\u003Efrom\u003C\u002Fspan\u003E \u003Cspan class=\&nn\&\u003E__future__\u003C\u002Fspan\u003E \u003Cspan class=\&kn\&\u003Eimport\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eprint_function\u003C\u002Fspan\u003E\n\n\u003Cspan class=\&kn\&\u003Eimport\u003C\u002Fspan\u003E \u003Cspan class=\&nn\&\u003Etensorflow\u003C\u002Fspan\u003E \u003Cspan class=\&kn\&\u003Eas\u003C\u002Fspan\u003E \u003Cspan class=\&nn\&\u003Etf\u003C\u002Fspan\u003E\n\n\u003Cspan class=\&c1\&\u003E# Import MNIST data\u003C\u002Fspan\u003E\n\u003Cspan class=\&kn\&\u003Efrom\u003C\u002Fspan\u003E \u003Cspan class=\&nn\&\u003Etensorflow.examples.tutorials.mnist\u003C\u002Fspan\u003E \u003Cspan class=\&kn\&\u003Eimport\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Einput_data\u003C\u002Fspan\u003E\n\n\u003Cspan class=\&c1\&\u003E# Import itchat & threading\u003C\u002Fspan\u003E\n\u003Cspan class=\&kn\&\u003Eimport\u003C\u002Fspan\u003E \u003Cspan class=\&nn\&\u003Eitchat\u003C\u002Fspan\u003E\n\u003Cspan class=\&kn\&\u003Eimport\u003C\u002Fspan\u003E \u003Cspan class=\&nn\&\u003Ethreading\u003C\u002Fspan\u003E\n\n\u003Cspan class=\&c1\&\u003E# Create a running status flag\u003C\u002Fspan\u003E\n\u003Cspan class=\&n\&\u003Elock\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ethreading\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003ELock\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E()\u003C\u002Fspan\u003E\n\u003Cspan class=\&n\&\u003Erunning\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&bp\&\u003EFalse\u003C\u002Fspan\u003E\n\n\u003Cspan class=\&c1\&\u003E# Parameters\u003C\u002Fspan\u003E\n\u003Cspan class=\&n\&\u003Elearning_rate\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&mf\&\u003E0.001\u003C\u002Fspan\u003E\n\u003Cspan class=\&n\&\u003Etraining_iters\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E3C\u002Fspan\u003E\n\u003Cspan class=\&n\&\u003Ebatch_size\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E128\u003C\u002Fspan\u003E\n\u003Cspan class=\&n\&\u003Edisplay_step\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E10\u003C\u002Fspan\u003E\n\n\u003Cspan class=\&k\&\u003Edef\u003C\u002Fspan\u003E \u003Cspan class=\&nf\&\u003Enn_train\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ewechat_name\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eparam\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E):\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Eglobal\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Elock\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Erunning\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# Lock\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Ewith\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Elock\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Erunning\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&bp\&\u003ETrue\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# mnist data reading\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Emnist\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Einput_data\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eread_data_sets\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&s2\&\u003E\&data\u002F\&\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eone_hot\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&bp\&\u003ETrue\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Parameters\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# learning_rate = 0.001\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# training_iters = 3C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# batch_size = 128\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# display_step = 10\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Elearning_rate\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etraining_iters\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebatch_size\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Edisplay_step\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eparam\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Network Parameters\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003En_input\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E784\u003C\u002Fspan\u003E \u003Cspan class=\&c1\&\u003E# MNIST data input (img shape: 28*28)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003En_classes\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E10\u003C\u002Fspan\u003E \u003Cspan class=\&c1\&\u003E# MNIST total classes (0-9 digits)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Edropout\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&mf\&\u003E0.75\u003C\u002Fspan\u003E \u003Cspan class=\&c1\&\u003E# Dropout, probability to keep units\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# tf Graph input\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eplaceholder\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Efloat32\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&bp\&\u003ENone\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003En_input\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ey\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eplaceholder\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Efloat32\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&bp\&\u003ENone\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003En_classes\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ekeep_prob\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eplaceholder\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Efloat32\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E \u003Cspan class=\&c1\&\u003E#dropout (keep probability)\u003C\u002Fspan\u003E\n\n\n
\u003Cspan class=\&c1\&\u003E# Create some wrappers for simplicity\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Edef\u003C\u002Fspan\u003E \u003Cspan class=\&nf\&\u003Econv2d\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003EW\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eb\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Estrides\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E):\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# Conv2D wrapper, with bias and relu activation\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Enn\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Econv2d\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003EW\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Estrides\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Estrides\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Estrides\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E],\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Epadding\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'SAME'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Enn\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ebias_add\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eb\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Ereturn\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Enn\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erelu\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n\n\n
\u003Cspan class=\&k\&\u003Edef\u003C\u002Fspan\u003E \u003Cspan class=\&nf\&\u003Emaxpool2d\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ek\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E2\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E):\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# MaxPool2D wrapper\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Ereturn\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Enn\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Emax_pool\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eksize\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ek\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ek\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E],\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Estrides\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ek\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ek\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E],\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Epadding\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'SAME'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n\n\n
\u003Cspan class=\&c1\&\u003E# Create model\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Edef\u003C\u002Fspan\u003E \u003Cspan class=\&nf\&\u003Econv_net\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eweights\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebiases\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Edropout\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E):\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# Reshape input picture\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ereshape\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eshape\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E-\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E28\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E28\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Convolution Layer\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Econv1\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Econv2d\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eweights\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'wc1'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E],\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebiases\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'bc1'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# Max Pooling (down-sampling)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Econv1\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Emaxpool2d\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Econv1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ek\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E2\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Convolution Layer\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Econv2\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Econv2d\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Econv1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eweights\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'wc2'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E],\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebiases\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'bc2'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# Max Pooling (down-sampling)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Econv2\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Emaxpool2d\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Econv2\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ek\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E2\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Fully connected layer\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# Reshape conv2 output to fit fully connected layer input\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Efc1\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ereshape\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Econv2\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E-\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eweights\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'wd1'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E]\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eget_shape\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E()\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eas_list\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E()[\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E0\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E]])\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Efc1\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eadd\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ematmul\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Efc1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eweights\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'wd1'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E]),\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebiases\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'bd1'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Efc1\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Enn\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erelu\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Efc1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# Apply Dropout\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Efc1\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Enn\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Edropout\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Efc1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Edropout\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Output, class prediction\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Eout\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eadd\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ematmul\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Efc1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eweights\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'out'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E]),\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebiases\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'out'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Ereturn\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eout\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Store layers weight & bias\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Eweights\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&p\&\u003E{\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# 5x5 conv, 1 input, 32 outputs\u003C\u002Fspan\u003E\n
\u003Cspan class=\&s1\&\u003E'wc1'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003EVariable\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erandom_normal\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E([\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E5\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E5\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E32\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])),\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# 5x5 conv, 32 inputs, 64 outputs\u003C\u002Fspan\u003E\n
\u003Cspan class=\&s1\&\u003E'wc2'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003EVariable\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erandom_normal\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E([\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E5\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E5\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E32\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E64\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])),\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# fully connected, 7*7*64 inputs, 1024 outputs\u003C\u002Fspan\u003E\n
\u003Cspan class=\&s1\&\u003E'wd1'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003EVariable\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erandom_normal\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E([\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E7\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E*\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E7\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E*\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E64\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003EC\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])),\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# 1024 inputs, 10 outputs (class prediction)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&s1\&\u003E'out'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003EVariable\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erandom_normal\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E([\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003EC\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003En_classes\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E]))\u003C\u002Fspan\u003E\n
\u003Cspan class=\&p\&\u003E}\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&n\&\u003Ebiases\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&p\&\u003E{\u003C\u002Fspan\u003E\n
\u003Cspan class=\&s1\&\u003E'bc1'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003EVariable\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erandom_normal\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E([\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E32\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])),\u003C\u002Fspan\u003E\n
\u003Cspan class=\&s1\&\u003E'bc2'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003EVariable\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erandom_normal\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E([\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E64\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])),\u003C\u002Fspan\u003E\n
\u003Cspan class=\&s1\&\u003E'bd1'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003EVariable\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erandom_normal\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E([\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003EC\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])),\u003C\u002Fspan\u003E\n
\u003Cspan class=\&s1\&\u003E'out'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003EVariable\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erandom_normal\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E([\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003En_classes\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E]))\u003C\u002Fspan\u003E\n
\u003Cspan class=\&p\&\u003E}\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Construct model\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Epred\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Econv_net\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eweights\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebiases\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ekeep_prob\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Define loss and optimizer\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ecost\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ereduce_mean\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Enn\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Esoftmax_cross_entropy_with_logits\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Elogits\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Epred\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Elabels\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ey\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E))\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Eoptimizer\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etrain\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003EAdamOptimizer\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Elearning_rate\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Elearning_rate\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eminimize\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ecost\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Evaluate model\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ecorrect_pred\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eequal\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eargmax\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Epred\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E),\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eargmax\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ey\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E))\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Eaccuracy\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ereduce_mean\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ecast\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ecorrect_pred\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Efloat32\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E))\u003C\u002Fspan\u003E\n\n\n
\u003Cspan class=\&c1\&\u003E# Initializing the variables\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Einit\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eglobal_variables_initializer\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E()\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Launch the graph\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Ewith\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etf\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003ESession\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E()\u003C\u002Fspan\u003E \u003Cspan class=\&k\&\u003Eas\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Esess\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Esess\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erun\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Einit\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Estep\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# Keep training until reach max iterations\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Eprint\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'Wait for lock'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Ewith\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Elock\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Erun_state\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Erunning\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Eprint\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'Start'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Ewhile\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Estep\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E*\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebatch_size\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E&\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etraining_iters\u003C\u002Fspan\u003E \u003Cspan class=\&ow\&\u003Eand\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Erun_state\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ebatch_x\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebatch_y\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Emnist\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etrain\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Enext_batch\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ebatch_size\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# Run optimization op (backprop)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Esess\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erun\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eoptimizer\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Efeed_dict\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E{\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebatch_x\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ey\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebatch_y\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ekeep_prob\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Edropout\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E})\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Eif\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Estep\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E%\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Edisplay_step\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E==\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E0\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E\n
\u003Cspan class=\&c1\&\u003E# Calculate batch loss and accuracy\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Eloss\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eacc\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Esess\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erun\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E([\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ecost\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eaccuracy\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E],\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Efeed_dict\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E{\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebatch_x\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ey\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebatch_y\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ekeep_prob\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&mf\&\u003E1.\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E})\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Eprint\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&s2\&\u003E\&Iter \&\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E+\u003C\u002Fspan\u003E \u003Cspan class=\&nb\&\u003Estr\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Estep\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E*\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ebatch_size\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E+\u003C\u002Fspan\u003E \u003Cspan class=\&s2\&\u003E\&, Minibatch Loss= \&\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E+\u003C\u002Fspan\u003E \\\n
\u003Cspan class=\&s2\&\u003E\&{:.6f}\&\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eformat\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eloss\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E+\u003C\u002Fspan\u003E \u003Cspan class=\&s2\&\u003E\&, Training Accuracy= \&\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E+\u003C\u002Fspan\u003E \\\n
\u003Cspan class=\&s2\&\u003E\&{:.5f}\&\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eformat\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eacc\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E))\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Eitchat\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Esend\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&s2\&\u003E\&Iter \&\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E+\u003C\u002Fspan\u003E \u003Cspan class=\&nb\&\u003Estr\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Estep\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E*\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ebatch_size\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E+\u003C\u002Fspan\u003E \u003Cspan class=\&s2\&\u003E\&, Minibatch Loss= \&\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E+\u003C\u002Fspan\u003E \\\n
\u003Cspan class=\&s2\&\u003E\&{:.6f}\&\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eformat\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eloss\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E+\u003C\u002Fspan\u003E \u003Cspan class=\&s2\&\u003E\&, Training Accuracy= \&\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E+\u003C\u002Fspan\u003E \\\n
\u003Cspan class=\&s2\&\u003E\&{:.5f}\&\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eformat\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eacc\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E),\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ewechat_name\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Estep\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E+=\u003C\u002Fspan\u003E \u003Cspan class=\&mi\&\u003E1\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Ewith\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Elock\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Erun_state\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Erunning\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Eprint\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&s2\&\u003E\&Optimization Finished!\&\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Eitchat\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Esend\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&s2\&\u003E\&Optimization Finished!\&\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ewechat_name\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&c1\&\u003E# Calculate accuracy for 256 mnist test images\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Eprint\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&s2\&\u003E\&Testing Accuracy:\&\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \\\n
\u003Cspan class=\&n\&\u003Esess\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erun\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eaccuracy\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Efeed_dict\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E{\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Emnist\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etest\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eimages\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[:\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E256\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E],\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ey\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Emnist\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etest\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Elabels\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[:\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E256\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E],\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ekeep_prob\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&mf\&\u003E1.\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E}))\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Eitchat\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Esend\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&s2\&\u003E\&Testing Accuracy: \u003C\u002Fspan\u003E\u003Cspan class=\&si\&\u003E%s\u003C\u002Fspan\u003E\u003Cspan class=\&s2\&\u003E\&\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E%\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Esess\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Erun\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eaccuracy\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Efeed_dict\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E{\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Ex\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Emnist\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etest\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eimages\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[:\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E256\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E],\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ey\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Emnist\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etest\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Elabels\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[:\u003C\u002Fspan\u003E\u003Cspan class=\&mi\&\u003E256\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E],\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ekeep_prob\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E \u003Cspan class=\&mf\&\u003E1.\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E}),\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ewechat_name\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n\n
\u003Cspan class=\&k\&\u003Ewith\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Elock\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Erunning\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&bp\&\u003EFalse\u003C\u002Fspan\u003E\n\n\u003Cspan class=\&nd\&\u003E@itchat.msg_register\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E([\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Eitchat\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Econtent\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003ETEXT\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E])\u003C\u002Fspan\u003E\n\u003Cspan class=\&k\&\u003Edef\u003C\u002Fspan\u003E \u003Cspan class=\&nf\&\u003Echat_trigger\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Emsg\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E):\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Eglobal\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Elock\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Erunning\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Elearning_rate\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etraining_iters\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebatch_size\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Edisplay_step\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Eif\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Emsg\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'Text'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E]\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E==\u003C\u002Fspan\u003E \u003Cspan class=\&s1\&\u003Eu'开始'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Eprint\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'Starting'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Ewith\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Elock\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Erun_state\u003C\u002Fspan\u003E \u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Erunning\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Eif\u003C\u002Fspan\u003E \u003Cspan class=\&ow\&\u003Enot\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Erun_state\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\&\u003Etry\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E:\u003C\u002Fspan\u003E\n
\u003Cspan class=\&n\&\u003Ethreading\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003EThread\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Etarget\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Enn_train\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Eargs\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E=\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Emsg\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E[\u003C\u002Fspan\u003E\u003Cspan class=\&s1\&\u003E'FromUserName'\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E],\u003C\u002Fspan\u003E \u003Cspan class=\&p\&\u003E(\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Elearning_rate\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Etraining_iters\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Ebatch_size\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E,\u003C\u002Fspan\u003E \u003Cspan class=\&n\&\u003Edisplay_step\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E)))\u003C\u002Fspan\u003E\u003Cspan class=\&o\&\u003E.\u003C\u002Fspan\u003E\u003Cspan class=\&n\&\u003Estart\u003C\u002Fspan\u003E\u003Cspan class=\&p\&\u003E()\u003C\u002Fspan\u003E\n
\u003Cspan class=\&k\