Skip to content

Conversation

@jackzhxng
Copy link
Collaborator

@jackzhxng jackzhxng commented Dec 9, 2025

Requires the following upstream change and corresponding pin bump - huggingface/transformers#42822

Currently stuck at quant ops not getting delegated to xnnpack cc @metascroy

>           raise RuntimeError(f"Missing out variants: {missing_out_vars}")
E           RuntimeError: Missing out variants: {'torchao::quantize_affine', 'torchao::choose_qparams_affine', 'torchao::dequantize_affine'}

@jackzhxng jackzhxng force-pushed the jz/ministral-3 branch 2 times, most recently from 7745158 to b888fd9 Compare December 11, 2025 17:53
@jackzhxng jackzhxng force-pushed the jz/bump-transformers branch from 82b9ee3 to f334032 Compare December 11, 2025 17:56
@jackzhxng jackzhxng force-pushed the jz/bump-transformers branch from f334032 to 8a7b706 Compare December 11, 2025 17:56
@jackzhxng jackzhxng marked this pull request as draft December 11, 2025 18:16
@jackzhxng jackzhxng changed the base branch from jz/bump-transformers to jz/bump-transformers-2 December 11, 2025 19:22
@jackzhxng jackzhxng force-pushed the jz/bump-transformers-2 branch from 2feaf57 to 0154636 Compare December 12, 2025 02:03
@jackzhxng jackzhxng force-pushed the jz/ministral-3 branch 2 times, most recently from 4cd28a5 to 7f646a5 Compare December 12, 2025 02:52
@jackzhxng jackzhxng changed the base branch from jz/bump-transformers-2 to jz/bump-transformers December 12, 2025 02:52
@jackzhxng jackzhxng marked this pull request as ready for review December 12, 2025 18:30
Base automatically changed from jz/bump-transformers to main December 12, 2025 22:45
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants