Skip to content

PTQ-Level Efficiency, QAT-Level Accuracy | 训练后量化的效率,量化感知训练的精度

Commits on main

auto round quantizer supports gptq kernel (#155)
wenhuach21 committed 2 years ago
9cae103
fix qbits issue (#153)
wenhuach21 committed 2 years ago
c313fa3
Qbits related log (#151)
Wang, Zhe committed 2 years ago
34274fb
autoround_support_qbits_backend (#145)
Wang, Zhe committed 2 years ago
dbdc4a3
fix incorrect setting for lm-head (#149)
wenhuach21 committed 2 years ago
9da2bee
fix triton issue (#148)
wenhuach21 committed 2 years ago
04ea869
refine the code (#143)
wenhuach21 committed 2 years ago
59c6402
Fix exlllamav2 backend issue (#144)
wenhuach21 committed 2 years ago
e614d13
794cd90
fix typos (#140)
WeiweiZhang1 committed 2 years ago
4d2d259
bump version into v0.2 (#139)
chen, suyue committed 2 years ago
aafb82e
fix export issue with torch 2.0 (#129)
wenhuach21 committed 2 years ago
5bff86e
Update falcon recipe (#128)
wenhuach21 committed 2 years ago
416ec7e
e2985fd
Update phi2 recipe (#124)
wenhuach21 committed 2 years ago
edca298
17024b1
update gemma recipe (#121)
wenhuach21 committed 2 years ago
ecc1dd6
Remove unused hook (#122)
Sun, Xuehao committed 2 years ago
c7751c4
support `transformers.Conv1D` packing (#118)
Kaihui-intel committed 2 years ago
02a6660
Fix export format issue (#120)
wenhuach21 committed 2 years ago
ed29cf5
Fix doc (#119)
wenhuach21 committed 2 years ago
7bc9d8f
b51cfa9
fix lm-head gradient accumulation bug (#113)
wenhuach21 committed 2 years ago
ecca534
update shells (#112)
WeiweiZhang1 committed 2 years ago
3c214db
20% speedup by removing new zero tensor (#110)
wenhuach21 committed 2 years ago
c7434b6
remove costly operations
wenhuach21 committed 2 years ago
ad3a7bb