Repository navigation
[executorch][native] Cover negative, rounded and subnormal GGUF super-block scales in the op samples - #24543
Open
pytorchbot wants to merge 1 commit into
Open
[executorch][native] Cover negative, rounded and subnormal GGUF super-block scales in the op samples#24543pytorchbot wants to merge 1 commit into
pytorchbot wants to merge 1 commit into
Conversation
…-block scales in the op samples Pull Request resolved: #24368 GGUF samples used fixed positive power-of-two super-block scales, so engines' decoders never saw negative, rounded or subnormal ones. A model's i-th `nn.Linear` now takes the i-th scale family, and a new `scale_families` case gives each family its own output. A zero-d family waits behind a TODO. With q4_k's d at 0 every weight is -dmin·m, so a token's 32 outputs share one activation-quantization error (correlation about 0.74) and `OutputCheck`'s bias standard error, which assumes independent outputs, comes out about 4.9 times too small; with q6_k's d at 0 every weight is zero and the outputs are just the bias. Authored with Claude Code. @bypass-github-export-checks @bypass-github-pytorch-ci-checks @bypass-github-executorch-ci-checks ghstack-source-id: 463657756 @exported-using-ghexport Differential Revision: [D128254896](https://our.internmc.facebook.com/intern/diff/D128254896/)
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/24543
Note: Links to docs will display an error until the docs builds have been completed. ❗ 1 Active SEVsThere are 1 currently active SEVs. If your PR is affected, please view them below: ❌ 2 New Failures, 1 Cancelled Job, 1 Unrelated FailureAs of commit cea68c8 with merge base 0f8e096 ( NEW FAILURES - The following jobs have failed:
CANCELLED JOB - The following job was cancelled. Please retry:
FLAKY - The following job failed but was likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
meta-cla
Bot
added
the
CLA Signed
This branch was previously deployed
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
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 was created by the merge bot to help merge the original PR into the main branch.
ghstack PR number: #24368 by @digantdesai
^ Please use this as the source of truth for the PR details, comments, and reviews
ghstack PR base: https://github.com/pytorch/executorch/tree/gh/digantdesai/195/base
ghstack PR head: https://github.com/pytorch/executorch/tree/gh/digantdesai/195/head
Merge bot PR base: https://github.com/pytorch/executorch/tree/gh/digantdesai/194/orig
Merge bot PR head: https://github.com/pytorch/executorch/tree/gh/digantdesai/195/orig
@diff-train-skip-merge