[GitHub] [incubator-tvm] FrozenGene commented on issue #5230: Adding support for TFLite QnnSubtract operator.

[GitBox <gi...@apache.org>]

FrozenGene commented on issue #5230: Adding support for TFLite QnnSubtract operator.
URL: https://github.com/apache/incubator-tvm/pull/5230#issuecomment-610737656
 
 
   > @FrozenGene If we have C = 1, then depth wise conv becomes normal conv. There is nothing to accumulate across input channels basically. And depth_multiplier becomes equal to the number of output channels. What do you think? Is the change good with you?
   
   I think it is ok. I met this thing ever. That is when input channel is  1 (for example, gray image), we will get depthwise convolution of multiplier greater than 1 rather than normal convolution.  Would you mind doing one performance comparison? I expect we could get better performance when we use normal convolution.

----------------------------------------------------------------
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:
users@infra.apache.org


With regards,
Apache Git Services