Simulated W4Afp8 Quantization #331
wenhuach21
merged
35 commits into
main
from
fp8
try to support fp8
23e6ff42
add files
fb17014f
![pre-commit-ci[bot]](https://avatars.githubusercontent.com/in/68672?v=4&s=60)
[pre-commit.ci] auto fixes from pre-commit.com hooks
35cd3762
Merge branch 'main' into fp8
6bb17160
tiny change
97d1237c
fix
6d0c5828
fix nan issue and change to dynamic per token
fe41d5af
Merge branch 'main' into fp8
53ccc1c2
Merge branch 'fp8' of https://github.com/intel/auto-round into fp8
cf6f7199
support static quantization, the code is ugly
fe385c3f
fix
73ebb24f
refine a little
2b647808
update a little
237f886d
refine code, fp16 model are easily gen NAN grad, need to have a study
bd8fea4e
Merge branch 'main' into fp8
cba9bd20
tmp change
d9074475
fix a critic bug
23d96043
refine code
b5902596
merge conv1d and fix conv1d exporting issue
8ba1137d
Merge branch 'main' into fp8
dcaee165
tmp change
bebe8f93
Merge branch 'main' into fp8
41a7eab9
fix issue
9c111878
update
ac44576d
Merge branch 'main' into fp8
0c98a2e6
refine a little
2c190ae4
Merge branch 'fp8' of https://github.com/intel/auto-round into fp8
2e095b8d
fix preci issue
b0683228
fix preci issue
80a74ae6
remove debug code
07acfd69
try to fix ut
41ad46cc
Merge branch 'main' into fp8
cd18f37c
WeiweiZhang1
approved these changes
on 2024-11-28
n1ck-guo
approved these changes
on 2024-11-28
fix numba pack
5a94ac6b
Merge branch 'main' into fp8
2ad66ca0
fix comment
1608c978
Reviewers
Assignees
No one assigned
Labels
None yet
Milestone
No milestone
Login to write a write a comment.
Login via GitHub