Lunderberg commented on PR #13936:
URL: https://github.com/apache/tvm/pull/13936#issuecomment-1482808030

   @Hzfengsy , thank you on the reminder, and it would be good to get this PR 
finished and merged in.
   
   @junrushao Could I get your assistance on the last CI error?  The failing 
unit tests are in 
[`test_cuda_c2d`](https://github.com/apache/tvm/blob/main/tests/python/unittest/test_meta_schedule_space_cuda.py#L122),
 added in https://github.com/apache/tvm/pull/12043, but I'm not familiar enough 
with the meta-schedule search space generation to tell if this is a spurious 
failure.
   
   It looks like this simplification allowed the search space to go down an 
entirely different route than the unit test was expecting, as the generated 
sketch outputs the transpose (`conv2d_transpose_nhwc: T.Buffer((1, 8, 8, 
256))`) with a completely different shape than the expected output defined in 
the unit test (`conv2d_nhwc: T.Buffer((1, 112, 112, 64))`).


-- 
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.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to