[GitHub] Godricly commented on issue #7703: what is the meaning of dtype parameter in get_symbol() function?

[gi...@git.apache.org]

Godricly commented on issue #7703: what is the meaning of dtype parameter in get_symbol() function?
URL: https://github.com/apache/incubator-mxnet/issues/7703#issuecomment-326553972
 
 
   You don't have to since the float32 is the default setup. In most cases, we use float type in training. And the image is normalized into range [0,1]. float32 here means typical single precision float type, while float16 represents half precision float point. It requires less memory on GPU and runs faster for some Nvidia GPUs with half precision units. However, due to the loss of precision, the performance could be worse than single precision one.
 
----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services