meixitu opened a new issue #15425: in mxnet1.4.1-cuda10.0, depthwise conv  
training is very very slow
URL: https://github.com/apache/incubator-mxnet/issues/15425
 
 
   Note: Providing complete information in the most concise form is the best 
way to get help. This issue template serves as the checklist for essential 
information to most of the technical issues and bug reports. For non-technical 
issues and feature requests, feel free to present the information in what you 
believe is the best form.
   
   For Q & A and discussion, please start a discussion thread at 
https://discuss.mxnet.io 
   
   ## Description
   Depthwise conv is very slow in python3.6, mxnet1.4.1-cuda10.0 version.
   If I set num_group=1, the training speed can improve 20 times.
   
   
   ## Environment info (Required)
   
![image](https://user-images.githubusercontent.com/32910309/60470843-d9467e00-9c16-11e9-8aab-7c86d677db5c.png)
   
![image](https://user-images.githubusercontent.com/32910309/60470857-e5cad680-9c16-11e9-85fb-153fc7a303ea.png)
   
   
   
   
   ## Steps to reproduce
   (Paste the commands you ran that produced the error.)
   
   0, python=3.6.0
   1. pip3 install mxnet-cu100==1.4.1
   2. use this model, 
https://github.com/mnikitin/EfficientNet/blob/master/efficientnet_model.py
   3. set the batch_size=128, 4GPU, efficientnet-b6, input_size=112x112x3
   
   ## What have you tried to solve it?
   
   1. I try to use  mx.sym.Convolution to replace the gluon, the training speed 
is same
   
   

----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to 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