说明

• ?代码重现笔记
• ? renference book: 【Hands-On Machine Learning with Scikit-Learn & TensorFlow 】
• ?Chapter 9: 【Up and Running with TensorFlow】

准备工作 ?

• ? 配置深度学习环境
• 导入需用到package

import tensorflow as tf
import os
import numpy as np
from sklearn.datasets import fetch_california_housingos.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'        # 去除AVX2警告

Session的四种用法 ?

• Creating Your First Graph and Running It in a Session

def session_usage():print("tensorflow version:", tf.__version__)# the following code does not actually perform any computation.# it just creates a computation graph.x = tf.Variable(3, name="x")y = tf.Variable(4, name="y")f = x*x*y + y + 2sess = tf.Session()sess.run(x.initializer)sess.run(y.initializer)result = sess.run(f)print("第一种方式运行结果：", result)sess.close()with tf.Session() as sess_2:x.initializer.run()y.initializer.run()result = f.eval()       # 等价于 result = tf.get_default_session().run(f)print("第二种方式运行结果：", result)init = tf.global_variables_initializer()    # prepare an init nodewith tf.Session() as sess_3:init.run()result = f.eval()print("第三种方式运行结果：", result)# Inside Jupyter or within Python shellsess_4 = tf.InteractiveSession()init.run()result = f.eval()print("第四种方式运行结果：", result)sess_4.close()session_usage()

节点值的生命周期 ?

• In single-process TensorFlow, multiple sessions do not share any state, even if they reuse the same graph (each session would have its own copy of every variable). In distributed TensorFlow, variable state is stored on the servers, not in the sessions, so multiple sessions can share the same variables.

def lifecycle_of_node_value():w = tf.constant(3)x = w + 2y = x + 5z = x * 3# In short, the preceding code evaluates w and x twice.with tf.Session() as sess_1:print(y.eval())print(z.eval())# A variable starts its life when its initializer is run,# and it ends when the session is closed.# evaluate both y and z in just one graph runwith tf.Session() as sess_2:y_val, z_val = sess_2.run([y, z])print(y_val)print(z_val)lifecycle_of_node_value()

线性回归 | 正规方程法 ?

• Use Normal Equation ?
• Note that housing.target is a 1D array, but we need to reshape it to a column vector to compute theta.Recall that NumPy’s reshape() function accepts –1 (meaning “unspecified”) for one of the dimensions: that dimension will be computed based on the array’s length and the remaining dimensions.

housing = fetch_california_housing()
m, n = housing.data.shape       # 20640 * 8
housing_data_plus_bias = np.c_[np.ones((m, 1)), housing.data]   # 20640 * 9def normal_equation_linear_regression():X = tf.constant(housing_data_plus_bias, dtype=tf.float32, name="X")y = tf.constant(housing.target.reshape(-1, 1), dtype=tf.float32, name="y")XT = tf.transpose(X)# the Normal Equation: (θ = XT · X)^(–1) · XT · y;theta = tf.matmul(tf.matmul(tf.matrix_inverse(tf.matmul(XT, X)), XT), y)with tf.Session() as sess:theta_value = theta.eval()print(theta_value)normal_equation_linear_regression()

线性回归 | 批梯度下降法 ?

• Implementing Gradient Descent Using Batch Gradient Descent ?
• When using Gradient Descent, remember that it is important to first normalize the input feature vectors, or else training may be much slower. You can do this using TensorFlow, NumPy, Scikit-Learn’s StandardScaler, or any other solution you prefer.

def batch_gradient_descent_linear_regression():from sklearn.preprocessing import StandardScalern_epochs = 1000learning_rate = 0.01scaler = StandardScaler()scaled_housing_data = scaler.fit_transform(housing.data)scaled_housing_data_plus_bias = np.c_[np.ones((m, 1)), scaled_housing_data]X = tf.constant(scaled_housing_data_plus_bias, dtype=tf.float32, name="X")y = tf.constant(housing.target.reshape(-1, 1), dtype=tf.float32, name="y")theta = tf.Variable(tf.random_uniform([n + 1, 1], -1.0, 1.0), name="theta")y_pred = tf.matmul(X, theta, name="predictions")error = y_pred - ymse = tf.reduce_mean(tf.square(error), name="mse")# gradients = 2/m * tf.matmul(tf.transpose(X), error)         # derived from the cost functiongradients = tf.gradients(mse, [theta])[0]                     # using auto_diff to compute gradients automatically# implement the Batch Gradient Descent step θ(next step) = θ –η∇θMSE(θ)training_op = tf.assign(theta, theta - learning_rate * gradients)init = tf.global_variables_initializer()with tf.Session() as sess:sess.run(init)for epoch in range(n_epochs):if epoch % 100 == 0:print("Epoch", epoch, "MSE =", mse.eval())sess.run(training_op)best_theta = theta.eval()print("Best theta: {:}".format([theta[0] for theta in best_theta]))# batch_gradient_descent_linear_regression()

线性回归 | 优化器法 ?

• Using an Optimizer with Batch Gradient Descent

def optimizer_gradient_descent_linear_regression():from sklearn.preprocessing import StandardScalern_epochs = 1000learning_rate = 0.01scaler = StandardScaler()scaled_housing_data = scaler.fit_transform(housing.data)scaled_housing_data_plus_bias = np.c_[np.ones((m, 1)), scaled_housing_data]X = tf.constant(scaled_housing_data_plus_bias, dtype=tf.float32, name="X")y = tf.constant(housing.target.reshape(-1, 1), dtype=tf.float32, name="y")theta = tf.Variable(tf.random_uniform([n + 1, 1], -1.0, 1.0), name="theta")y_pred = tf.matmul(X, theta, name="predictions")error = y_pred - ymse = tf.reduce_mean(tf.square(error), name="mse")optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)# using another optimizer# optimizer = tf.train.MomentumOptimizer(learning_rate=learning_rate, momentum=0.9)training_op = optimizer.minimize(mse)init = tf.global_variables_initializer()with tf.Session() as sess:sess.run(init)for epoch in range(n_epochs):if epoch % 100 == 0:print("Epoch", epoch, "MSE =", mse.eval())sess.run(training_op)best_theta = theta.eval()print("Best theta: {:}".format([theta[0] for theta in best_theta]))optimizer_gradient_descent_linear_regression()

线性回归 | mini-批梯度下降法 ?

• Using an Optimizer with Mini-batch Gradient Descent
• we need a way to replace X and y at every iteration with the next mini-batch. The simplest way to do this is to use placeholder nodes. These nodes are special because they don’t actually perform any computation, they just output the data you tell them to output at runtime.They are typically used to pass the training data to TensorFlow during training. If you don’t specify a value at runtime for a placeholder, you get an exception.

def mini_batch_gradient_descent_linear_regression():from sklearn.preprocessing import StandardScalern_epochs = 1000learning_rate = 0.01scaler = StandardScaler()scaled_housing_data = scaler.fit_transform(housing.data)scaled_housing_data_plus_bias = np.c_[np.ones((m, 1)), scaled_housing_data]X = tf.placeholder(tf.float32, shape=(None, n+1), name="X")y = tf.placeholder(tf.float32, shape=(None, 1))theta = tf.Variable(tf.random_uniform([n + 1, 1], -1.0, 1.0, seed=42), name="theta")y_pred = tf.matmul(X, theta, name="predictions")error = y_pred - ymse = tf.reduce_mean(tf.square(error), name="mse")optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)# using another optimizer# optimizer = tf.train.MomentumOptimizer(learning_rate=learning_rate, momentum=0.9)training_op = optimizer.minimize(mse)init = tf.global_variables_initializer()n_epochs = 10  # smaller for mini-batchbatch_size = 100n_batchs = int(np.ceil(m / batch_size))def fetch_batch(epoch, batch_index, batch_size):np.random.seed(epoch * n_batchs + batch_index)indices = np.random.randint(m, size=batch_size)X_batch = scaled_housing_data_plus_bias[indices]y_batch = housing.target.reshape(-1, 1)[indices]return X_batch, y_batchwith tf.Session() as sess:sess.run(init)for epoch in range(n_epochs):for batch_index in range(n_batchs):X_batch, y_batch = fetch_batch(epoch, batch_index, batch_size)sess.run(training_op, feed_dict={X: X_batch, y: y_batch})best_theta = theta.eval()print("Best theta: {:}".format([theta[0] for theta in best_theta]))mini_batch_gradient_descent_linear_regression()

保存及加载模型 ?

• Saving and Restoring Models

def save_and_restore_model():tf.reset_default_graph()from sklearn.preprocessing import StandardScalern_epochs = 1000learning_rate = 0.01scaler = StandardScaler()scaled_housing_data = scaler.fit_transform(housing.data)scaled_housing_data_plus_bias = np.c_[np.ones((m, 1)), scaled_housing_data]X = tf.constant(scaled_housing_data_plus_bias, dtype=tf.float32, name="X")y = tf.constant(housing.target.reshape(-1, 1), dtype=tf.float32, name="y")theta = tf.Variable(tf.random_uniform([n + 1, 1], -1.0, 1.0, seed=42), name="theta")y_pred = tf.matmul(X, theta, name="predictions")error = y_pred - ymse = tf.reduce_mean(tf.square(error), name="mse")optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)training_op = optimizer.minimize(mse)# save modelinit = tf.global_variables_initializer()saver = tf.train.Saver()# for restoring theta with a different name, such as "weights"# saver = tf.train.Saver({"weight": theta})with tf.Session() as sess:sess.run(init)for epoch in range(n_epochs):if epoch % 100 == 0:print("Epoch", epoch, "MSE =", mse.eval())saver.save(sess, save_path="model_saved/my_model.ckpt")sess.run(training_op)best_theta = theta.eval()print("Best theta: {:}".format([theta[0] for theta in best_theta]))saver.save(sess, save_path="model_saved/my_model_final.ckpt")# restore modeltf.reset_default_graph()            # notice that we start with an empty graph# load the graph structuresaver = tf.train.import_meta_graph("model_saved/my_model_final.ckpt.meta")theta = tf.get_default_graph().get_tensor_by_name("theta:0")with tf.Session() as sess:saver.restore(sess, "model_saved/my_model_final.ckpt")best_theta_restored = theta.eval()print("比较两个值是否相近: ", np.allclose(best_theta, best_theta_restored))# save_and_restore_model()

可视化 ?

• Visualizing the Graph and Training Curves
• Using TensorBoard

def log_mse_for_visualization():from sklearn.preprocessing import StandardScalerfrom datetime import datetimetf.reset_default_graph()n_epochs = 1000learning_rate = 0.01scaler = StandardScaler()scaled_housing_data = scaler.fit_transform(housing.data)scaled_housing_data_plus_bias = np.c_[np.ones((m, 1)), scaled_housing_data]X = tf.placeholder(tf.float32, shape=(None, n+1), name="X")y = tf.placeholder(tf.float32, shape=(None, 1))theta = tf.Variable(tf.random_uniform([n + 1, 1], -1.0, 1.0, seed=42), name="theta")y_pred = tf.matmul(X, theta, name="predictions")error = y_pred - ymse = tf.reduce_mean(tf.square(error), name="mse")optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)# using another optimizer# optimizer = tf.train.MomentumOptimizer(learning_rate=learning_rate, momentum=0.9)training_op = optimizer.minimize(mse)init = tf.global_variables_initializer()# added for lognow = datetime.utcnow().strftime("%Y%m%d%H%M%S")root_logdir = "tf_logs"logdir = "{}/run-{}".format(root_logdir, now)mse_summary = tf.summary.scalar("MSE", mse)file_writer = tf.summary.FileWriter(logdir, tf.get_default_graph())n_epochs = 10  # smaller for mini-batchbatch_size = 100n_batchs = int(np.ceil(m / batch_size))def fetch_batch(epoch, batch_index, batch_size):np.random.seed(epoch * n_batchs + batch_index)indices = np.random.randint(m, size=batch_size)X_batch = scaled_housing_data_plus_bias[indices]y_batch = housing.target.reshape(-1, 1)[indices]return X_batch, y_batchwith tf.Session() as sess:sess.run(init)for epoch in range(n_epochs):for batch_index in range(n_batchs):X_batch, y_batch = fetch_batch(epoch, batch_index, batch_size)# added for logif batch_index % 10 == 0:summary_str = mse_summary.eval(feed_dict={X: X_batch, y: y_batch})step = epoch * n_batchs + batch_indexfile_writer.add_summary(summary_str, step)sess.run(training_op, feed_dict={X: X_batch, y: y_batch})file_writer.close()best_theta = theta.eval()print("Best theta: {:}".format([theta[0] for theta in best_theta]))log_mse_for_visualization()
"""
# cmd中输入以下命令，查看可视化结果
activate py36_tf
# cd 到tf_logs的上级目录
cd xxxxx/
tensorboard --logdir tf_logs/
# 复制出现的网址，用Chrome浏览器打开
"""

Tips | 名称作用域 ?

• Using Name scopes

def name_scope_for_tensorflow_skills():from sklearn.preprocessing import StandardScalerfrom datetime import datetimetf.reset_default_graph()n_epochs = 1000learning_rate = 0.01scaler = StandardScaler()scaled_housing_data = scaler.fit_transform(housing.data)scaled_housing_data_plus_bias = np.c_[np.ones((m, 1)), scaled_housing_data]X = tf.placeholder(tf.float32, shape=(None, n+1), name="X")y = tf.placeholder(tf.float32, shape=(None, 1))theta = tf.Variable(tf.random_uniform([n + 1, 1], -1.0, 1.0, seed=42), name="theta")y_pred = tf.matmul(X, theta, name="predictions")# added for name scope skillwith tf.name_scope("loss") as scope:error = y_pred - ymse = tf.reduce_mean(tf.square(error), name="mse")optimizer = tf.train.GradientDescentOptimizer(learning_rate=learning_rate)training_op = optimizer.minimize(mse)init = tf.global_variables_initializer()# added for lognow = datetime.utcnow().strftime("%Y%m%d%H%M%S")root_logdir = "tf_logs"logdir = "{}/run-{}".format(root_logdir, now)mse_summary = tf.summary.scalar("MSE", mse)file_writer = tf.summary.FileWriter(logdir, tf.get_default_graph())n_epochs = 10  # smaller for mini-batchbatch_size = 100n_batchs = int(np.ceil(m / batch_size))def fetch_batch(epoch, batch_index, batch_size):np.random.seed(epoch * n_batchs + batch_index)indices = np.random.randint(m, size=batch_size)X_batch = scaled_housing_data_plus_bias[indices]y_batch = housing.target.reshape(-1, 1)[indices]return X_batch, y_batchwith tf.Session() as sess:sess.run(init)for epoch in range(n_epochs):for batch_index in range(n_batchs):X_batch, y_batch = fetch_batch(epoch, batch_index, batch_size)# added for logif batch_index % 10 == 0:summary_str = mse_summary.eval(feed_dict={X: X_batch, y: y_batch})step = epoch * n_batchs + batch_indexfile_writer.add_summary(summary_str, step)sess.run(training_op, feed_dict={X: X_batch, y: y_batch})file_writer.flush()file_writer.close()best_theta = theta.eval()print("Best theta: {:}".format([theta[0] for theta in best_theta]))# name_scope_for_tensorflow_skills()

Tips | 常见的模块化 ?

• Using Common Modularity to build multiple ReLUs

def common_modularity_for_tensorflow_skills():tf.reset_default_graph()def relu(X):w_shape = (int(X.get_shape()[1]), 1)w = tf.Variable(tf.random_normal(w_shape), name="weights")b = tf.Variable(0.0, name="bias")z = tf.add(tf.matmul(X, w), b, name="z")return tf.maximum(z, 0., name="relu")n_feature = 3X = tf.placeholder(tf.float32, shape=(None, n_feature), name="X")relus = [relu(X) for i in range(5)]output = tf.add_n(relus, name="output")print("common modularity coding test: {:}".format([relu.op.name for relu in relus]))common_modularity_for_tensorflow_skills()

Tips | 更好的模块化 ?

• Using Better Modularity to build multiple ReLUs
• Note that when you create a node, TensorFlow checks whether its name already exists, and if it does it appends an underscore followed by an index to make the name unique.

def better_modularity_for_tensorflow_skills():tf.reset_default_graph()def relu(X):# added for better modularitywith tf.name_scope("relu"):w_shape = (int(X.get_shape()[1]), 1)w = tf.Variable(tf.random_normal(w_shape), name="weights")b = tf.Variable(0.0, name="bias")z = tf.add(tf.matmul(X, w), b, name="z")return tf.maximum(z, 0., name="max")n_feature = 3X = tf.placeholder(tf.float32, shape=(None, n_feature), name="X")relus = [relu(X) for i in range(5)]output = tf.add_n(relus, name="output")print("better modularity coding test: {:}".format([relu.op.name for relu in relus]))better_modularity_for_tensorflow_skills()

Tips | 基于参数的共享变量 ?

• Sharing Variables: Using Parameter

def using_parameter_sharing_variables_for_tensorflow_skills():tf.reset_default_graph()def relu(X, threshold):with tf.name_scope("relu"):w_shape = (int(X.get_shape()[1]), 1)w = tf.Variable(tf.random_normal(w_shape), name="weights")b = tf.Variable(0.0, name="bias")z = tf.add(tf.matmul(X, w), b, name="z")return tf.maximum(z, threshold, name="max")threshold = tf.Variable(0.0, name="threshold")n_feature = 3X = tf.placeholder(tf.float32, shape=(None, n_feature), name="X")relus = [relu(X, threshold) for i in range(5)]output = tf.add_n(relus, name="output")print("classic sharing variables using parameter coding test: {:}".format(output))using_parameter_sharing_variables_for_tensorflow_skills()

Tips | 基于函数属性的共享变量 ?

• Sharing Variables: Using Function Attribute

def using_attribute_sharing_variables_for_tensorflow_skills():tf.reset_default_graph()def relu(X):with tf.name_scope("relu"):if not hasattr(relu, "threshold"):relu.threshold = tf.Variable(0.0, name="threshold")w_shape = (int(X.get_shape()[1]), 1)w = tf.Variable(tf.random_normal(w_shape), name="weights")b = tf.Variable(0.0, name="bias")z = tf.add(tf.matmul(X, w), b, name="z")return tf.maximum(z, relu.threshold, name="max")n_feature = 3X = tf.placeholder(tf.float32, shape=(None, n_feature), name="X")relus = [relu(X) for i in range(5)]output = tf.add_n(relus, name="output")print("sharing variables using function attribute coding test: {:}".format(output))using_attribute_sharing_variables_for_tensorflow_skills()

Tips | 基于函数的共享变量 ?

• Sharing Variables: Using Function Provided By Tensorflow
• The idea is to use the get_variable() function to create the shared variable if it does not exist yet, or reuse it if it already exists. The desired behavior (creating or reusing) is controlled by an attribute of the current variable_scope().

def using_function_sharing_variables_for_tensorflow_skills():tf.reset_default_graph()def relu(X):with tf.variable_scope("relu", reuse=True):     # reuse 由get_variable()创建的threshold变量threshold = tf.get_variable("threshold")w_shape = (int(X.get_shape()[1]), 1)w = tf.Variable(tf.random_normal(w_shape), name="weights")b = tf.Variable(0.0, name="bias")z = tf.add(tf.matmul(X, w), b, name="z")return tf.maximum(z, threshold, name="max")n_feature = 3X = tf.placeholder(tf.float32, shape=(None, n_feature), name="X")with tf.variable_scope("relu"):                     # 第一次定义threshold变量threshold = tf.get_variable("threshold", shape=(),initializer=tf.constant_initializer(0.0))relus = [relu(X) for relu_index in range(5)]output = tf.add_n(relus, name="output")file_witer = tf.summary.FileWriter("logs/relu9", tf.get_default_graph())file_witer.close()using_function_sharing_variables_for_tensorflow_skills()

Tips | 基于函数的共享变量（升级版） ?

• Sharing Variables: Using Function Provided By Tensorflow
• solve the first call problem
• The resulting graph is slightly different than before, since the shared variable lives within the first ReLU

def using_better_function_sharing_variables_for_tensorflow_skills():tf.reset_default_graph()def relu(X):threshold = tf.get_variable("threshold", shape=(),initializer=tf.constant_initializer(0.0))w_shape = (int(X.get_shape()[1]), 1)w = tf.Variable(tf.random_normal(w_shape), name="weights")b = tf.Variable(0.0, name="bias")z = tf.add(tf.matmul(X, w), b, name="z")return tf.maximum(z, threshold, name="max")n_feature = 3X = tf.placeholder(tf.float32, shape=(None, n_feature), name="X")relus = []for relu_index in range(5):with tf.variable_scope("relu", reuse=(relu_index >= 1)) as scope:relus.append(relu(X))output = tf.add_n(relus, name="output")file_witer = tf.summary.FileWriter("logs/relu9_modified", tf.get_default_graph())file_witer.close()using_better_function_sharing_variables_for_tensorflow_skills()

tensorflow | 练习 ✒

• Tensorflow Skills Exercise
• variable name scope exercise

def variable_scope_exercise():tf.reset_default_graph()with tf.variable_scope("my_scope"):x0 = tf.get_variable("x", shape=(), initializer=tf.constant_initializer(0.))x1 = tf.Variable(0., name="x")x2 = tf.Variable(0., name="x")with tf.variable_scope("my_scope", reuse=True):x3 = tf.get_variable("x")x4 = tf.Variable(0., name="x")with tf.variable_scope("", default_name="", reuse=True):x5 = tf.get_variable("my_scope/x")print("x0:", x0.op.name)        # x0: my_scope/xprint("x1:", x1.op.name)        # x1: my_scope/x_1print("x2:", x2.op.name)        # x2: my_scope/x_2print("x3:", x3.op.name)        # x3: my_scope/xprint("x4:", x4.op.name)        # x4: my_scope_1/xprint("x5:", x5.op.name)        # x5: my_scope/xprint(x0 is x3 and x3 is x5)    # Truevariable_scope_exercise()

The End ??

• Congratulations?

初识 ❤ TensorFlow |【一见倾心】相关推荐

1. Tensorflow |（1）初识Tensorflow

   Tensorflow |(1)初识Tensorflow 关于 TensorFlow TensorFlow是一个采用数据流图(data flow graphs),用于数值计算的开源软件库.节点(Node ...

2. 【深度学习】初识tensorflow之分布式训练

   情况一.单机单卡 单机单卡是最普通的情况,当然也是最简单的,示例代码如下: #coding=utf-8 #单机单卡 #对于单机单卡,可以把参数和计算都定义再gpu上,不过如果参数模型比较大,显存不足等 ...

3. 初识 TensorFlow 旅程之一

   Table of Contents Table of Contents 一.TensorFlow 相关常用的数据处理 1.test_first 2.Matrix Operator 3.约简 (redu ...

4. 深度学习——初识TensorFlow

   书籍:<深度学习--基于Python语言和TensorFlow平台> 三好学生问题的引入: 学校要评选三好学生,根据德育分.智育分和体育分3项分数来计算一个总分,然后根据总分来确定谁能够被 ...

5. Tensorflow |（5）模型保存与恢复、自定义命令行参数

   Tensorflow |(1)初识Tensorflow Tensorflow |(2)张量的阶和数据类型及张量操作 Tensorflow |(3)变量的的创建.初始化.保存和加载 Tensorflow ...

6. Tensorflow |（4）名称域、图 和会话

   Tensorflow |(1)初识Tensorflow Tensorflow |(2)张量的阶和数据类型及张量操作 Tensorflow |(3)变量的的创建.初始化.保存和加载 Tensorflow ...

7. Tensorflow |（3）变量的的创建、初始化、保存和加载

   Tensorflow |(1)初识Tensorflow Tensorflow |(2)张量的阶和数据类型及张量操作 Tensorflow |(3)变量的的创建.初始化.保存和加载 变量的的创建.初始化 ...

8. Tensorflow |（2）张量的阶和数据类型及张量操作

   Tensorflow |(1)初识Tensorflow Tensorflow |(2)张量的阶和数据类型及张量操作 张量的阶和数据类型 TensorFlow用张量这种数据结构来表示所有的数据.你可以把 ...

9. 独家 | 手把手教TensorFlow（附代码）

   上一期我们发布了"一文读懂TensorFlow(附代码.学习资料)",带领大家对TensorFlow进行了全面了解,并分享了入门所需的网站.图书.视频等资料,本期文章就来带你一步步 ...

最新文章

1. Mozilla两款火狐插件包含恶意代码被紧急喊停
2. Oracle 启动，建表
3. silverlight(一.安装和新建项目)
4. SpringBoot无法找到加载类 ，应用Feign其他服务无法package两类问题
5. win7局域网计算机 慢,Win7系统开机宽带连接很慢怎么办？
6. python 枚举类型_Python枚举类
7. 百万数据php7取出循环_Thinkphp5 分批批量导出百万条数据记录的Code，不用PHPEXCEL哦！...
8. FT计算机系统,芯片CP/FT测试的基本概念理解
9. 飞腾服务器自带的kvm管理,领存技术飞腾FT2000+ 2U 12盘位存储服务器
10. 一文解决样本不均衡（全）
11. 个推的appid是指什么_私人教练要注意胸肌更快增厚的4个方法
12. php 编译 sass,如何在Symfony 3中使用纯PHP编译SASS(scss)
13. Linux/Unix文件系统索引节点浅析
14. 虚拟机装xp 架设服务器,VMware安装虚拟机系统ghost xp sp3步骤图解
15. 【转】阿里巴巴技术专家杨晓明：基于Hadoop技术进行地理空间分析
16. echart 地图 某个地区_一个让echarts中国地图包含省市轮廓的技巧
17. httpwatch使用_使用PHP自动化HTTPWatch
18. Java 解析epub格式电子书，helloWorld程序，附带源程序和相关jar包
19. 赢在互联网的起跑线上
20. iPadmini能运行c语言吗,一个移动固态硬盘，可以给iPad Pro2020和iPad Mini用是什么体验？...

热门文章

1. Docker推送一个自制镜像到dockerhub
2. Java网课基础笔记（25）19-08-07
3. python key lambda_python – 什么是key = lambda
4. ai领域职业规划_我如何抛弃我的咨询职业并进入技术领域
5. 半乳糖-人血清白蛋白 Gal-HSA，Gal-PEG-HSA 半乳糖修饰人血清白蛋白
6. vue3 前端pc生成微信支付二维码
7. csp-j2第四题解
8. 使用163邮箱发送邮件报错(554, b'DT:SPM 163 smtp3,G9xpCgCHi5RJOFVemMZ4Dw--.348S3 1582643274,please see http://ma
9. 【项目管理】项目进度管理
10. flutter　人脸识别_使用flutter和tensorflow lite进行人脸识别认证