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

refine md tables (#994)
Weiwei committed 1 year ago
c0aa30a
fix mllm device_map ut (#1000)
Tang Kaihui committed 1 year ago
4aeca3a
fix cuda ut bug (#999)
Heng Guo committed 1 year ago
d1bf7e8
284eecd
268f7dd
05cab09
add support for Magistral-Small (#980)
Heng Guo committed 1 year ago
dc6a6d3
fix guff scheme and device_map bug (#969)
Heng Guo committed 1 year ago
be146c7
update bits (#986)
Xin He committed 1 year ago
4afbe0a
b9245b5
refine readme (#978)
Wenhua Cheng committed 1 year ago
3f7bdac
06dd686
disable compile packing by default (#975)
Yi Liu committed 1 year ago
8c767db
fix cuda CI backend issue, fixtypo (#974)
Weiwei committed 1 year ago
635eded
merge main branch to alg_ext (#970)
Wenhua Cheng committed 1 year ago
9ecfaa5
support for model scope (#957)
Heng Guo committed 1 year ago
2448dcf
clean code (#968)
Wenhua Cheng committed 1 year ago
3c1a678
enhance flux doc (#967)
Wang, Mengni committed 1 year ago
7635f7e
fix rtn bug (#966)
Wang, Mengni committed 1 year ago
c4ef9a8
fix bug of imatrix contains 0 (#955)
Heng Guo committed 1 year ago
5e33cbc
fix critic disable_opt_rtn regression (#963)
Wenhua Cheng committed 1 year ago
282aab6
update readme (#962)
Wenhua Cheng committed 1 year ago
7d8016d
refine AutoScheme readme/code (#958)
Wenhua Cheng committed 1 year ago
8ac82a4
add logo (#960)
Wenhua Cheng committed 1 year ago
eb2facd
12c4984
Reduce AutoSchem VRAM usage by up to 10X (#944)
Wenhua Cheng committed 1 year ago
90c2fb4
fbb9c13