Trigger RuntimeError for when normalization is requested but not utilized
#76
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This change raises an error message for when the user has specified normalization arguments, but none of them were used. Signed-off-by: Lee, Kin Long Kelvin <kin.long.kelvin.lee@intel.com>
Also check that the number of used keys matches the number of keys passed by the user. Signed-off-by: Lee, Kin Long Kelvin <kin.long.kelvin.lee@intel.com>
laserkelvin
added
ux
|
I like the idea, but I'm not sure this actually does anything at the moment. The _make_normalizers function takes in whatever |
|
Okay that's good to know, I forgot that behavior. Let me rewrite this then! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This PR adds checks and informative error messages centered around target normalization.
BaseTaskModule._compute_losses, appending keys to aused_normlist as normalization is performed. When all the losses have been calculated, ifused_normis empty despite having requested normalization, we trigger aRuntimeErrorto stop training, and inform the user that no normalization was performed.@melo-gonzo I've made it trigger an error, rather than a warning message, since it's probably too easy to ignore warnings and rather have it complete training.
Closes #75