[GitHub] xingzhaoo opened a new issue #7313: Validation Accuracy does not grow and stay in constant, running alexnet on ImageNet1k

[git]

xingzhaoo opened a new issue #7313: Validation Accuracy does not grow and stay 
in constant, running alexnet on ImageNet1k 
URL: https://github.com/apache/incubator-mxnet/issues/7313
 
 
   I set exact same initial setting as alexnet paper suggested and ran alexnet 
on 1 machine in CPU only mode (kvstore='none'). The dataset is ImageNet1k 
(size=24G). But the val accuracy stays constant after 20 Epochs.   
   Does anyone have any suggestion why the val accuracy does not grow for 
Alexnet?
   
   
![image](https://user-images.githubusercontent.com/15624862/28884313-5df8a146-777f-11e7-8f8b-6720970a073c.png)
   
 
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on GitHub and use the
URL above to go to the specific comment.
 
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


With regards,
Apache Git Services