krea2: fix timestep-shift training#947
Open
tcldr wants to merge 1 commit into
Open
Conversation
tcldr
force-pushed
the
krea2-patch-size-fix
branch
from
July 14, 2026 20:58
976c1d8 to
04b6add
Compare
This file contains hidden or 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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
With
timestep_type: shift, krea2 trains on a far more high-noise-biased timestep distribution than intended becauseBaseSDTrainProcessis expecting to find the token patch size viapatch_sizerather than thepatchconfig key used by the Krea 2 types – so the lookup falls back to the incorrectpatch_size = 1Fix
Probably a few ways to do this, but figured renaming the config type from
patchtopatch_size– so that BaseSDTrainProcess can find the expected property – and updating all internal consumers would be cleanest.