Skip to content

Add quark as an autoquantizer#14077

Draft
amd-inechakr wants to merge 5 commits into
huggingface:mainfrom
amd-inechakr:add-quark-quantizer
Draft

Add quark as an autoquantizer#14077
amd-inechakr wants to merge 5 commits into
huggingface:mainfrom
amd-inechakr:add-quark-quantizer

Conversation

@amd-inechakr

Copy link
Copy Markdown

What does this PR do?

Fixes # (issue)

Before submitting

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

amd-inechakr and others added 5 commits June 23, 2026 18:59
Two pre-existing import bugs surfaced by end-to-end testing against an
amd-quark working tree:

  * src/diffusers/quantizers/quantization_config.py:789
  * tests/quantization/quark/test_quark.py:110

Both imported 'Config as QConfig' from quark.torch.quantization.config.config,
but that module only ever exported QConfig (no class named Config exists in
amd-quark, on any branch). Every from_pretrained on a quark-quantized
checkpoint failed with ImportError. Switch to importing QConfig directly.

Also update docs/source/en/quantization/quark.md to match amd-quark's
post-huggingface#5249 calibration API: get_calib_dataloader now yields plain dicts
keyed by forward parameter names, consumed via model(**data); the
WrappingModel adapter and .inner_model accessor are no longer exposed.
@github-actions github-actions Bot added documentation Improvements or additions to documentation quantization tests utils size/L PR with diff > 200 LOC labels Jun 26, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

documentation Improvements or additions to documentation quantization size/L PR with diff > 200 LOC tests utils

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant