You are viewing a plain text version of this content. The canonical link for it is here.

Posted to commits@tvm.apache.org by GitBox <gi...@apache.org> on 2019/11/11 03:43:01 UTC

[GitHub] [incubator-tvm] optima2005 opened a new pull request #4300: [Relay][Frontend][Tensorflow]Add conv2d_transpose

optima2005 opened a new pull request #4300: [Relay][Frontend][Tensorflow]Add conv2d_transpose
URL: https://github.com/apache/incubator-tvm/pull/4300
 
 
   Hi,  This PR is to fix #4263
   
   After investigation, I found this operation has been implemented in relay cc codes and been working in keras frontend. For tensorflow frontend, only python part is missing. So this PR only to fix the python part. 
   
   The conv2d_transpose implementation is similar with the conv2d op, beside the following differences,
   1) Its name in tensorflow graph definition is 'Conv2DBackpropInput'
   2) Its inputs order: 0 -- "conv2d_transpose/input_sizes", 1 -- weight,  2 -- input_placeholder
   3) It needs an extra input parameter "output_shape"
   
   The PR implementation is reusing _conv() function and adding the conditions to handle the below differences.

----------------------------------------------------------------
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