Fix inference issue in ARK for AWQ format #1170
luoyu-intel
merged
23 commits into
main
from
fix_1154
wenhuach21
commented
on 2025-12-19
wenhuach21
commented
on 2025-12-19
wenhuach21
commented
on 2025-12-19
wenhuach21
commented
on 2025-12-19
wenhuach21
approved these changes
on 2025-12-19
fix awq accuracy
e986f40f
use auto compute dtype for gptq model
fe1b399d
add awq UT cases
f1a39917
enable xpu+awq
28d53a0e
![pre-commit-ci[bot]](https://avatars.githubusercontent.com/in/68672?v=4&s=60)
[pre-commit.ci] auto fixes from pre-commit.com hooks
d785a544
rm assert for xpu
01510189
wenhuach21
commented
on 2025-12-19
rm assert
4cb28c6b
wenhuach21
commented
on 2025-12-22
update model test cases to pytest
89b0e00c
[pre-commit.ci] auto fixes from pre-commit.com hooks
69c903cd
add feature checker
ee957377
rm alignment
892ad331
add more bits for UT
fded35ad
[pre-commit.ci] auto fixes from pre-commit.com hooks
ea210d7d
Update auto_round_extension/ark/qlinear.py
8533a9f7
Update auto_round_extension/ark/qlinear.py
11ef58ad
split params to reduce test combinations.
f60c41db
fix threshold for 2bit
ffd54832
[pre-commit.ci] auto fixes from pre-commit.com hooks
3046c4e8
merge all Qlinears into one
0abc9f4d
remove int8 from awq
21f6a97c
fix copilot suggestion
4f485b3c
[pre-commit.ci] auto fixes from pre-commit.com hooks
aa84f62f
Merge branch 'main' into fix_1154
d31a878a
Reviewers
Assignees
No one assigned
Labels
None yet
Milestone
Login to write a write a comment.
Login via GitHub