Skip to content

[executorch][native-ggml] Reject weight-only quantized linears unless activation quant is allowed - #24536

Closed
pytorchbot wants to merge 1 commit into
gh/digantdesai/187/origfrom
gh/digantdesai/188/orig
Closed

pytorchbot wants to merge 1 commit into
gh/digantdesai/187/origfrom
gh/digantdesai/188/orig

Conversation

@pytorchbot

Copy link
Copy Markdown
Collaborator

This PR was created by the merge bot to help merge the original PR into the main branch.
ghstack PR number: #24361 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/188/base
ghstack PR head: https://github.com/pytorch/executorch/tree/gh/digantdesai/188/head
Merge bot PR base: https://github.com/pytorch/executorch/tree/gh/digantdesai/187/orig
Merge bot PR head: https://github.com/pytorch/executorch/tree/gh/digantdesai/188/orig

@diff-train-skip-merge

… activation quant is allowed

Pull Request resolved: #24361

ggml quantizes the activations of its quantized matmuls, so weight-only int4 and GGUF linears silently ran with numerics the PTN doesn't describe. Compile now rejects them unless `GgmlEngineOptions::allow_activation_quant_for_weight_only` is set. `ggml_ptn_runner` and `run_hf_llm_ggml` expose it as `--allow_activation_quant_for_weight_only`, which GGUF LLMs on ggml now need.

Authored with Claude Code.

@bypass-github-export-checks
@bypass-github-pytorch-ci-checks
@bypass-github-executorch-ci-checks
ghstack-source-id: 463657742
@exported-using-ghexport

Differential Revision: [D128254887](https://our.internmc.facebook.com/intern/diff/D128254887/)
@pytorch-bot

pytorch-bot Bot commented Nov 3, 2026 •

Copy link
Copy Markdown

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/24536

Note: Links to docs will display an error until the docs builds have been completed.

❌ 1 New Failure

As of commit 77f2cad with merge base 0f8e096 (image):

NEW FAILURE - The following job has failed:

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@meta-cla meta-cla Bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Nov 3, 2026
@digantdesai

Copy link
Copy Markdown
Contributor

Landed on main via #24698.

@digantdesai digantdesai closed this Nov 5, 2026

This branch was previously deployed

1 inactive deployment
cadence — 77f2cad3 Deployed Nov 3, 2026 by pytorchbot
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants