测试训练>> mnist-bn>> 返回
项目作者： soloice

项目描述 ：
Using slim to perform batch normalization
高级语言： Python
项目地址: git://github.com/soloice/mnist-bn.git
创建时间： 2017-02-14T08:45:27Z
项目社区：https://github.com/soloice/mnist-bn
开源协议：
下载


mnist-bn

Using slim to perform batch normalization

Run python mnist_bn.py --phase=train to train.
Run python mnist_bn.py --phase=test to test.

It should achieve an accuracy of ~99.3% or higher on test set.

I’ve added accuracy, cross_entropy and batch normalization paramters into summary. Use tensorboard —logdir=/log to explore the learning curve and parameter distributions!