import tensorflow as tf

import numpy as np

# create data

x_data = np.random.rand(100).astype(np.float32)

y_data = x_data*0.1 + 0.3

#搭建模型

Weights = tf.Variable(tf.random_uniform([1], -1.0, 1.0))

biases = tf.Variable(tf.zeros([1]))

y = Weights*x_data + biases

#计算误差

loss = tf.reduce_mean(tf.square(y-y_data))

#传播误差

#反向传递误差的工作就教给 optimizer 了，我们使用的误差传递方法是梯度下降法: Gradient Descent 让后我们使用 optimizer 来进行参数的更新.

optimizer = tf.train.GradientDescentOptimizer(0.5)

train = optimizer.minimize(loss)

#训练

#到目前为止，我们只是建立了神经网络的结构，还没有使用这个结构。在使用这个结构之前，我们必须先初始化所有之前定义的 Variable, 所以这一步是很重要的！

# init = tf.initialize_all_variables () # tf 马上就要废弃这种写法

init = tf.global_variables_initializer()  # 替换成这样就好

#接着，我们再创建会话 Session. 我们用 Session 来执行 init 初始化步骤。并且，用 Session 来 run 每一次 training 的数据。逐步提升神经网络的预测准确性.

sess = tf.Session()

sess.run(init)          # Very important

for step in range(201):

    sess.run(train)

    if step % 20 == 0:

        print(step, sess.run(Weights), sess.run(biases))

import tensorflow as tf

import numpy as np

#creat data

x_data = np.random.rand(100).astype(np.float32)

y_data = x_data*01.+0.3

### creat tensorflow structure start ###

Weights = tf.Variable(tf.random.uniform((1,),-1.0,1.0))

biases = tf.Variable(tf.zeros([1]))

y = Weights*x_data+biases

loss= lambda: tf.keras.losses.MSE(y_data,Weights * x_data +biases)

optimizer = tf.keras.optimizers.SGD(learning_rate=0.5)

### creat tensorflow structure end ###

for step in range(201):

    optimizer.minimize(loss, var_list=[Weights, biases])

    if step %20==0:

        #print(step,sess.run(Weights),sess.run(biases))

        print("{} step, weights = {}, biases = {}".format(step, Weights.read_value(), biases.read_value()))  # read_value 函数可用 numpy 替换

import tensorflow as tf

# create two matrixes

matrix1 = tf.constant([[3,3]])

matrix2 = tf.constant([[2],

                       [2]])

product = tf.matmul(matrix1,matrix2)

# method 1

sess = tf.Session()

result = sess.run(product)

print(result)

sess.close()

# [[12]]

# method 2

with tf.Session() as sess:

    result2 = sess.run(product)

    print(result2)

# [[12]]

import tensorflow as tf

matrix1=tf.constant([[3,3]])

matrix2=tf.constant([[2],[2]])

product=tf.matmul(matrix1,matrix2)

tf.print(product)

import tensorflow as tf

state = tf.Variable(0, name='counter')

# 定义常量 one

one = tf.constant(1)

# 定义加法步骤 (注：此步并没有直接计算)

new_value = tf.add(state, one)

# 将 State 更新成 new_value

update = tf.assign(state, new_value)

# 如果定义 Variable, 就一定要 initialize

init = tf.global_variables_initializer()  # 替换成这样就好

# 使用 Session

with tf.Session() as sess:

    sess.run(init)

    for _ in range(3):

        sess.run(update)

        print(sess.run(state))

import tensorflow as tf

state=tf.Variable(0,name='counter')

one=tf.constant(1)

new_value=tf.add(state,one)

for _ in range(3):

	#The variable value can be changed using one of the assign methods.

	state.assign_add(new_value)

	tf.print(state,new_value)

import tensorflow as tf

#在 Tensorflow 中需要定义 placeholder 的 type ，一般为 float32 形式

input1 = tf.placeholder(tf.float32)

input2 = tf.placeholder(tf.float32)

# mul = multiply 是将 input1 和 input2 做乘法运算，并输出为 output 

ouput = tf.multiply(input1, input2)

with tf.Session() as sess:

    print(sess.run(ouput, feed_dict={input1: [7.], input2: [2.]}))

# [ 14.]

%config IPCompleter.greedy=True  #按 tag 可以自动补全代码

import tensorflow as tf

from tensorflow import keras

import matplotlib.pyplot as plt

#导入数据集

fasion_mnist = keras.datasets.fashion_mnist

#划分数据集

(train_images,train_labels),(test_images,test_labels) = fasion_mnist.load_data()

#看数据集大小

print(train_images.shape)

#展示图片

plt.imshow(train_images[1])

model = keras.Sequential()

model.add(keras.layers.Flatten(input_shape=(28,28)))

model.add(keras.layers.Dense(128,activation=tf.nn.relu))

model.add(keras.layers.Dense(10,activation=tf.nn.softmax))

model.summary()

#以下为 console

Model: "sequential_1"

_________________________________________________________________

Layer (type)                 Output Shape              Param #   

=================================================================

flatten_1 (Flatten)          (None, 784)  个像素             0         

_________________________________________________________________

dense (Dense)                (None, 128)               100480    （784个像素+1权重）*128个神经元

_________________________________________________________________

dense_1 (Dense)              (None, 10)                1290      （128个神经元+1权重）*10个类别

=================================================================

Total params: 101,770

Trainable params: 101,770

Non-trainable params: 0

_________________________________________________________________

#scaling or normalization 把灰度变成只有 1 和 0

train_imges_scaled=train_images/255

#metrics=['accuracy'] 是为了看到指标 accuracy 的变化，

model.compile(optimizer=tf.optimizers.Adam(),loss=tf.losses.sparse_categorical_crossentropy,metrics=['accuracy'])

#训练，epochs 重复次数 5 次

model.fit(train_imges_scaled,train_labels,epochs=5)

#模型评估

test_images_scaled = test_images/255

model.evaluate(test_images_scaled,test_labels)

import numpy as np

print(np.argmax( model.predict([[test_images[0]/255]])))

print(test_labels[0])

plt.imshow(test_images[0])

#重载一个 callback 方法

class myCallback(tf.keras.callbacks.Callback):

    def on_epoch_end(self,epoch,logs={}):

        if(logs.get('loss')<0.4):

            print("\n Loss is low so cancelling training!")

            self.model.stop_training = True

callbacks= myCallback()

mnist =tf.keras.datasets.fashion_mnist

(training_images,training_labels),(test_images,test_labels) = mnist.load_data()

training_images_scaled = training_images/255

test_images_scaled = test_images/255

model =tf.keras.models.Sequential([

    tf.keras.layers.Flatten(),

    tf.keras.layers.Dense(512,activation = tf.nn.relu),

    tf.keras.layers.Dense(10,activation = tf.nn.softmax)

])

model.compile(optimizer='adam',loss='sparse_categorical_crossentropy',metrics=['accuracy'])

model.fit(training_images_scaled,training_labels,epochs=5,callbacks=[callbacks])

---console:---

Epoch 1/5

1875/1875 [==============================] - 5s 3ms/step - loss: 0.4734 - accuracy: 0.8292

Epoch 2/5

1867/1875 [============================>.] - ETA: 0s - loss: 0.3585 - accuracy: 0.8697

 Loss is low so cancelling training!

1875/1875 [==============================] - 5s 3ms/step - loss: 0.3583 - accuracy: 0.8697