PR #1699 support model_free WOQ quantization

support model_free WOQ quantization #1699

xin3he merged 51 commits into main from xinhe/4-14

implement model free

dc592e99

polished implementation

177bf48b

remove useless gpu_concurrency

97e03620

添加预编译模式匹配器以提高量化过程中的性能和可扩展性

ff47a97a

fix typo

4d9ad0e5

update document

58709e64

remove useless code and update UT

d3951f26

mend

16991ea2

remove high_gpu_mem_usage since no performacen benefit.

83b9b4fe

update regex

687260db

fix bug and simplify UT

68d0cb7b

fix bug

312f75df

add WOQ limiation and support bits group_size setting

3ca4d3b5

xin3he requested a review from

copilot-pull-request-reviewer 64 days ago

Merge branch 'main' into xinhe/4-14

3f15e02d

[pre-commit.ci] auto fixes from pre-commit.com hooks

47b3f35d

update doc

76f99151

minor fix

c588ad22

copilot-pull-request-reviewer commented on 2026-04-17

enable quant_nontext_module

0c141653

xin3he requested a review from

changwangss 61 days ago

xin3he requested a review from

yiliu30 61 days ago

xin3he requested a review from

wenhuach21 61 days ago

xin3he requested a review from

WeiweiZhang1 61 days ago

xin3he requested a review from

n1ck-guo 61 days ago

n1ck-guo commented on 2026-04-21

yiliu30 commented on 2026-04-21

Enhance model-free quantization support and improve documentation

405de53d

Merge remote-tracking branch 'origin/main' into xinhe/4-14

6c5ce29a

support loading pytorch_model.bin and ignore conv1d embed by creating…

0697324a

add UT to cover conv1d detection

f4fc5f41

support MXFP4/8 dequantization

4f6f97e4

Merge branch 'main' into xinhe/4-14

ed46cd68

fix pylint

7e3a3f87

Merge branch 'main' into xinhe/4-14

958191a5

add auto fallback and change class name

7440c321

xin3he requested a review from

n1ck-guo 56 days ago

xin3he requested a review from

yiliu30 56 days ago

fix CI

8b8d084e

wenhuach21 commented on 2026-04-26

update readme

eb5fdf43

添加回退压缩器功能以支持量化和保存

98a50401

Merge branch 'main' into xinhe/4-14

46465c39

support diffusion model

7c76188a

fix bug

a92acc2b

support layer_config={".ffn.experts.": {"scheme": "W2A16"}} usage

46ed32c4

fix bug

6f41cec5

update UT

9f81c67c

fix bug

16ead43b

Merge remote-tracking branch 'origin/main' into xinhe/4-14

48994a40

add model free for new arch

3d9812ca

[pre-commit.ci] auto fixes from pre-commit.com hooks

bd318616

Merge branch 'main' into xinhe/4-14

efd8753a

wenhuach21 approved these changes on 2026-04-30

fix issue in comments

312eabef

unify cli content

dbdaf9f1

[pre-commit.ci] auto fixes from pre-commit.com hooks

6ba73d61

update per comments

bba56fe8

Merge branch 'main' into xinhe/4-14

fca22106

fix bug

22711c39

Merge branch 'main' into xinhe/4-14

7e537ef7

fix CI

7aac7952

remove breakpoint

1888f86c

xin3he force pushed from f593cfa2 to 551e1ca5 42 days ago

xin3he force pushed from 551e1ca5 to f9cd4c7d 42 days ago

add iters in init kwargs for new arch

f9cd4c7d

xin3he merged f0013f09 into main 40 days ago

xin3he deleted the xinhe/4-14 branch 40 days ago

Reviewers

wenhuach21

yiliu30

n1ck-guo

copilot-pull-request-reviewer

changwangss

WeiweiZhang1

Assignees

No one assigned

Labels

None yet

Milestone

No milestone

auto-round support model_free WOQ quantization #1699 Merged

support model_free WOQ quantization #1699

auto-round
support model_free WOQ quantization
#1699

Merged