Repository navigation
Fix GGUF tensor alignment and validate model config - #20
Closed
dipeshbabu wants to merge 3 commits into
Closed
dipeshbabu wants to merge 3 commits into
dipeshbabu wants to merge 3 commits into
Conversation
Author
|
This also fixes for avoiding per-token malloc/leak by using mmap string views. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
GGUF loading now aligns tensor data for the declared alignment and rejects unsupported layer counts, missing feed-forward dimensions, and incompatible attention/KV-head ratios before inference. Missing KV-head metadata defaults to ordinary multi-head attention.
Attention accumulation uses the existing output buffer, allowing head dimensions above 256 without a fixed-size stack array or another allocation. Failed loads release the mapped file.
Validation:
make -C picolm nativepassed on Linux. Synthetic GGUF load/forward tests passed under AddressSanitizer and UndefinedBehaviorSanitizer, covering two alignments, a 512-wide head, and four invalid dimension configurations. No model download was needed. The upstream Build PicoLLM workflow currently requires maintainer approval.