[GitHub] [tvm] comaniac edited a comment on pull request #7975: Remove warning which is adding too much noise

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

comaniac edited a comment on pull request #7975:
URL: https://github.com/apache/tvm/pull/7975#issuecomment-832870770


   In summary:
   
   (A) by @masahi: Turn the warning to DEBUG so that only the advance developers would see it by purpose. We could of course add a paragraph of how to enable this message in the tutorial.
   
   (B) by @tkonolige: Improve the message to make it more informative. This is also a valuable direction so I try to make it more concrete. First, we remove the last scary sentence from the message:
   ```
   WARNING autotvm:compile_engine.py:256 Cannot find config for target=llvm -keys=cpu -link-params=0, workload=('conv2d_NCHWc.x86', ('TENSOR', (1, 1, 32, 32), 'float32'), ('TENSOR', (8, 1, 5, 5), 'float32'), (1, 1), (0, 0, 0, 0), (1, 1), 'NCHW', 'NCHW', 'float32').
   ```
   Second, when we are about to display this warning for the first time, we also display the message as Tristan mentioned:
   ```
   One or more config are missing. Please tune your model for better performance (link).
   ```
   
   (C) by myself: Combine both options. That is, we only show the overall warning once and put the detail schedule missing warnings to DEBUG.
   ```
   One or more config are missing. Please tune your model for better performance (link), or use DEBUG logging level to see more details.
   ```
   
   @rkimball @mbrookhart @masahi @tkonolige what do you folks think?


-- 
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