Refactor regularization step closes #117 #200
Merged
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
SebChw
force-pushed
the
regularize_changes
branch
from
41be9dd to
7a29414
Compare
SebChw
force-pushed
the
regularize_changes
branch
from
7a29414 to
af127da
Compare
kordc
reviewed
Dec 2, 2023
kordc
approved these changes
Dec 2, 2023
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Outdated suggestions cannot be applied.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.
To see this step in action please check
get rid off turn_on/off regularization - We won't use it for regularization
add datamodule modifiers - regularization will be based on these modifiers
reduce dependencies - since we use them just once they are not needed
add access to current stage for ArtModule - During regularization when augmentations are used I must use them only during training
add utils for orange logs - I made some warning orange
add utils for building metric name - Small util usefull when we want to define ModelCallback
add possibility for the step to decide when it should be run or not - This is a big deal as for me - new regularizations should be continued even though the previous one failed to obtain good accuracy. That's why selected steps can decide if they should be run or not. I know this is not with SOLID but I couldn't find any other way
separate model and data logs - When log_params were executed on dataloader withouth
setupthere was a crush.rewrite regularizaton - Rewritten step
add class fields - Decide which steps can decide for themselves
adjust dashboard - Since regularize steps adds new field to parameters dashboard must handle it