chinakook commented on issue #7613: 1x1 convolution acceleration
URL: https://github.com/apache/incubator-mxnet/pull/7613#issuecomment-325193542
 
 
   @reminisce I will complete the testings. I think performance improvement and 
memory reduction will show up in the image detection and segmentation cases 
where feature maps are very big. You can refer to the tensorflow's optimization 
for the 1x1 conv, it's more simple and more clear than Caffe:
   [tensorflow conv 
forward](https://github.com/tensorflow/tensorflow/blob/a0d784bdd31b27e013a7eac58a86ba62e86db299/tensorflow/core/kernels/conv_ops_using_gemm.cc#L238)
   [tensorflow conv 
backward](https://github.com/tensorflow/tensorflow/blob/42ca99b5aae03a8122ba0db94abfe1f3f5c257dc/tensorflow/core/kernels/conv_grad_input_ops.cc#L666)
 
----------------------------------------------------------------
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

Reply via email to