tensorflow2.0--实现前向传播

import tensorflow as tf
from tensorflow import keras
from tensorflow.keras import datasets
import os

调整输出信息，只输出error

os.environ[‘TF_CPP_MIN_LOG_LEVEL’]=‘2’

获取数据

(x,y),_=datasets.mnist.load_data()

把x值从0-255调整为0-1 把np转换为tensor

x=tf.convert_to_tensor(x,dtype=tf.float32)/255.
y=tf.convert_to_tensor(y,dtype=tf.int32)
print(x.shape,y.shape,x.dtype,y.dtype)
print(tf.reduce_min(x),tf.reduce_max(x))
print(tf.reduce_min(y),tf.reduce_max(y))

设置一批数据为128

train_db=tf.data.Dataset.from_tensor_slices((x,y)).batch(128)
train_iter=iter(train_db)
sample=next(train_iter)
print(‘batch:’,sample[0].shape,sample[1].shape)

定义参数与学习率

w1=tf.Variable(tf.random.truncated_normal([784,256],stddev=0.1))
b1=tf.Variable(tf.zeros([256]))
w2=tf.Variable(tf.random.truncated_normal([256,128],stddev=0.1))
b2=tf.Variable(tf.zeros([128]))
w3=tf.Variable(tf.random.truncated_normal([128,10],stddev=0.1))
b3=tf.Variable(tf.zeros([10]))
lr=1e-3

循环数据集10次

for epoch in range(10):

#分批循环数据集
for step,(x,y) in enumerate(train_db):
    #把X转换为[batch,784]
    x=tf.reshape(x,[-1,28*28])
    #tensor提供的自动求导
    with tf.GradientTape() as tape:
        #x[128,784]->[128,256]
        h1=x@w1+tf.broadcast_to(b1,[x.shape[0],256])
        #非线性激励
        h1=tf.nn.relu(h1)
        #x[128,256]->[128,128]
        h2=h1@w2+b2
        h2=tf.nn.relu(h2)
        #输出[128,10]
        out=h2@w3+b3
        #对y进行one_hot转变为[128,10]
        y_onehot=tf.one_hot(y,depth=10)
        #损失函数：(y-y')**2/128
        loss=tf.square(y_onehot-out)
        loss=tf.reduce_mean(loss)
    #传入损失函数，参数
    grads=tape.gradient(loss,[w1,b1,w2,b2,w3,b3])
    #更新参数
    w1.assign_sub(lr*grads[0])
    b1.assign_sub(lr*grads[1])
    w2.assign_sub(lr*grads[2])
    b2.assign_sub(lr*grads[3])
    w3.assign_sub(lr*grads[4])
    b3.assign_sub(lr*grads[5])
    #输出损失值
    if step%100==0:
        print(epoch,step,'loss',float(loss))