comaniac commented on a change in pull request #6612:
URL: https://github.com/apache/incubator-tvm/pull/6612#discussion_r498579481
##########
File path: python/tvm/autotvm/tuner/tuner.py
##########
@@ -174,6 +176,13 @@ def tune(self, n_trial, measure_option,
early_stopping=None, callbacks=(), si_pr
else:
logger.setLevel(old_level)
+ if error_ct == i:
+ logging.warning(
+ "Could not find a single schedule for task %s. A fallback
config will be used. "
+ "The following errors occured:\n %s",
+ self.task,
+ "\n ".join(errors),
Review comment:
This could be super long if you set trial to a large number (e.g.,
1,500) and it could easily wipe out the most important message (i.e., no valid
schedule found). How about dumping them to a log file and only show the log
file name here?
##########
File path: python/tvm/autotvm/tuner/tuner.py
##########
@@ -174,6 +176,13 @@ def tune(self, n_trial, measure_option,
early_stopping=None, callbacks=(), si_pr
else:
logger.setLevel(old_level)
+ if error_ct == i:
+ logging.warning(
+ "Could not find a single schedule for task %s. A fallback
config will be used. "
Review comment:
```suggestion
"Could not find any valid schedule for task %s."
```
IMHO, "A fallback config will be used" is not a proper message here and
could be removed, because AutoTVM doesn't really compile a model. When you are
really compiling the model, you will see the corresponding message.
----------------------------------------------------------------
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:
[email protected]
