strip padding from text latent cache for Ernie#1520
Merged
Conversation
Ernie's text encoder hidden state is truncated per-batch to the longest prompt (variable length, not a fixed sequence length like Flux2), so it benefits from the same PruneMaskedTokens/PadMaskedTokens pattern as QwenBaseDataLoader to keep the disk cache compact. Co-Authored-By: Claude Sonnet 4.6 <[email protected]>
dxqb
added a commit
that referenced
this pull request
Jun 14, 2026
dxqb
added a commit
to dxqb/OneTrainer
that referenced
this pull request
Jun 19, 2026
Mirrors Nerogar#1520 (Ernie): tokens_mask/text_encoder_hidden_state are fixed at PROMPT_MAX_LENGTH after EncodeLensText's crop, so caching wastes disk space on padding for shorter prompts. Prune before caching, pad back to PROMPT_MAX_LENGTH on load. Co-Authored-By: Claude Sonnet 4.6 <[email protected]>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
implements #1345 by @BitcrushedHeart for Ernie
I probably missed it because the Ernie implementation was based on Flux2 - and pruning isn't possible for Flux2. It is possible for Ernie
Test plan
pre-commit run --all-filespassesAI assistance