auto-round
feat: support LLMC/ModelOpt/Quark format NVFP4 -> MXFP4 conversion
#2492
Merged
Merged
feat: support LLMC/ModelOpt/Quark format NVFP4 -> MXFP4 conversion #2492
xin3he
merged
10 commits into
main
from
xinhe/10-14
feat: add NVFP2MXFP scheme for model-free quantization, converting NV…
b9d7ff5d
fix: update NVFP4 to MXFP4 conversion documentation for clarity and a…
be6d303a
wenhuach21
requested changes
on 2026-10-14
Refactor NVFP4 to MXFP4 conversion and update tests
19bd64fa
![pre-commit-ci[bot]](https://avatars.githubusercontent.com/in/68672?v=4&s=60)
[pre-commit.ci] auto fixes from pre-commit.com hooks
85c1f56d
Refactor NVFP4 to MXFP4 conversion logic and update documentation
1b87871f
feat: add explicit MXFP targets handling and update related tests
4eaa130a
xin3he
changed the title
feat: enhance MXFP4 support with routed experts handling and update r…
b4f91e05
feat: support 'auto_round:llm_compressor' format in NVFP4 to MXFP4 co…
8a269e8f
feat: implement finalize_llmcompressor_config_names utility and updat…
dda577f7
wenhuach21
commented
on 2026-10-16
wenhuach21
commented
on 2026-10-16
wenhuach21
commented
on 2026-10-16
wenhuach21
approved these changes
on 2026-10-16
feat: refactor dtype detection and update related functions and tests
cc1ba046
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Login to write a write a comment.
Login via GitHub