这里我们采用Mnist手写数字数据集来讲解如何将训练过程可视化。
https://github.com/MagaretJi/Tensorboard
Tensorboard记录与展示的数据形式:
(1)标量Scalars:tf.summary.scalar()
(2)图片Images:tf.summary.image()
(3)音频Audio:tf.summary.audio()
(4)计算图Graph:tf.summary.graph()
(5)数据分布Distribution:tf.summary.distribution()
(6)直方图Histograms:tf.summary.histogram()
(7)嵌入向量Embeddings
Step1
导入需要用到的包
我们直接使用了TensorFlow自带的数据集。
import tensorflow as tf
import numpy as np
from tensorflow.examples.tutorials.mnist import input_data
数据集准备
mnist=input_data.read_data_sets("MNIST_data/",one_hot=True)
Step2 设置占位符
x:训练集样本特征
y:样本分类标签
with tf.name_scope('input'):
x=tf.placeholder(tf.float32,[None,784],name='x-input')
y=tf.placeholder(tf.float32,[None,10],name='y-input')
Step3 设置参数概要
我们知道,在训练的过程在参数是不断地在改变和优化的,我们往往想知道每次迭代后参数都做了哪些变化,可以将参数的信息展现在tenorbord上,因此我们专门写一个方法来收录每次的参数信息。计算出Variable的mean,stddev,max和min。
# 参数概要
def variable_summaries(var):
with tf.name_scope('summaries'):
mean = tf.reduce_mean(var)
tf.summary.scalar('mean',mean)
with tf.name_scope('stddev'):
stddev = tf.sqrt(tf.reduce_mean(tf.square(var - mean)))
tf.summary.scalar('stddev',stddev)
tf.summary.scalar('max',tf.reduce_max(var))
tf.summary.scalar('min',tf.reduce_min(var))
tf.summary.histogram('histogram',var)
Step4 搭建神经网络
with tf.name_scope('layer'):
with tf.name_scope('weight'):
w =tf.Variable(tf.zeros([784,10]),name='weight')
variable_summaries(w)
with tf.name_scope('bias'):
b=tf.Variable(tf.zeros([10]),name='bias')
variable_summaries(b)
with tf.name_scope('softmax'):
prediction=tf.nn.softmax(tf.matmul(x,w)+b,name='softmax')
variable_summaries(prediction)
Step5
Loss值和梯度下降
tf.summary.scalar(tags, values, collections=None, name=None)
一般在画loss,accuary时会用到这个函数
with tf.name_scope('loss'):
loss=tf.reduce_mean(tf.square(y-prediction))
tf.summary.scalar('loss',loss)
with tf.name_scope('train'):
train_step=tf.train.GradientDescentOptimizer(0.2).minimize(loss)
准确率
with tf.name_scope('accuracy'):
#结果放在布尔型列表中
with tf.name_scope('correct_prediction'):
correct_prediction=tf.equal(tf.math.argmax(y,1),tf.math.argmax(prediction,1))
#tf.summary.histogram('accuracy/correct_prediction',correct_prediction)
with tf.name_scope('accuracy'):
#准确率
accuracy=tf.reduce_mean(tf.cast(correct_prediction,tf.float32))
tf.summary.histogram('accuracy/accuracy',accuracy)
Step6 超参数设置
#批次大小
batch_size=100
#批次数
n_batch=mnist.train.num_examples // batch_size
#训练轮数
n_epoch=210
Step7 开始训练
tf.summaries.merge_all()
合并默认图形中的所有汇总
tf.summary.FileWriter()
SummaryWriter是一个类,它可以调用以下成员函数来往event文件中添加相关的数据 addsummary(), add sessionlog(), add_event(), or add_graph()
with tf.Session() as sess:
sess.run(init)
merged = tf.summary.merge_all()
#日志存放路径,会存放所有汇总数据供Tensorflow展示
writer=tf.summary.FileWriter('./logs',sess.graph)
for epoch in range(n_epoch):
for batch in range(n_batch):
batch_xs,batch_ys=mnist.train.next_batch(batch_size)
summary,_=sess.run([merged,train_step],feed_dict={x:batch_xs,y:batch_ys})
writer.add_summary(summary,epoch)
acc = sess.run(accuracy,feed_dict={x:mnist.test.images,y:mnist.test.labels})
print("Lter " + str(epoch) + ",Testing Accuracy " + str(acc))
参考
https://blog.csdn.net/sinat_33761963/article/details/62433234