-
Notifications
You must be signed in to change notification settings - Fork 3.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Improve val step logging #7351
Improve val step logging #7351
Conversation
Codecov Report
@@ Coverage Diff @@
## master #7351 +/- ##
=======================================
Coverage 91% 92%
=======================================
Files 200 200
Lines 12925 12991 +66
=======================================
+ Hits 11801 11904 +103
+ Misses 1124 1087 -37 |
pytorch_lightning/trainer/connectors/logger_connector/logger_connector.py
Show resolved
Hide resolved
@awaelchli I think the functionality will stay the same, only the code will change. The screenshot you show is correct and gives the new behaviour following this PR. Val step logging is and odd usage anyway but at least this stops the potentially hundreds of graphs from being created in tensorboard. |
Okay, as long as all the other logging for training and epoch end is not affected. Consider changing the labelling here because these per-epoch plots being created was a design decision and therefore expected behaviour. |
could not agree more 🤣 |
pytorch_lightning/trainer/connectors/logger_connector/logger_connector.py
Outdated
Show resolved
Hide resolved
pytorch_lightning/trainer/connectors/logger_connector/logger_connector.py
Outdated
Show resolved
Hide resolved
pytorch_lightning/trainer/connectors/logger_connector/logger_connector.py
Show resolved
Hide resolved
…onnector.py Co-authored-by: ananthsub <ananth.subramaniam@gmail.com>
What does this PR do?
Fixes #7346
/epoch_*
to the metric names (which creates many plots in e.g. tensorboard)log_evaluation_step_metrics
to be in theLoggerConnector
(where it belongs haha)Before submitting
PR review
Anyone in the community is free to review the PR once the tests have passed.
Before you start reviewing make sure you have read Review guidelines. In short, see the following bullet-list:
Did you have fun?
Make sure you had fun coding 🙃