zhreshold commented on issue #9185: Gluon provided ResNet does not get 
desirable accuracy on CIFAR10
URL: 
https://github.com/apache/incubator-mxnet/issues/9185#issuecomment-357066788
 
 
   @fedorzh I have switched to ImageRecordIter with cifar_train.rec just like 
https://github.com/tornadomeet/ResNet/blob/master/symbol_resnet.py you 
mentioned, with gluon model, I got ~95% validation acc. So it's not the problem 
with gluon model, but the way to apply data augmentation.
   
   With that confirmed, using Dataloader with proper transformation can also 
achieve good accuracy IMO.

----------------------------------------------------------------
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:
[email protected]


With regards,
Apache Git Services

Reply via email to