intel/auto-round

Pull Requests Commits

yiliu30 committed 183 days ago

b82adadc

yiliu30 committed 183 days ago

47136ef7

remove result to cpu (#1115)

yiliu30 committed 184 days ago

Verified 5346cb73

fix quant fp8 model with iters=0 and scheme=nvfp4 (#1114)

n1ck-guo committed 185 days ago

Verified 99e4e75e

fix gguf packing device (#1105)

n1ck-guo committed 185 days ago

Verified 4f64730f

fix release version cuda ut fail (#1110)

n1ck-guo committed 185 days ago

Verified c9ac9a4d

fix auto-scheme/alg-ext multiple devices issue (#1107)

wenhuach21 committed 185 days ago

Verified 3e26b0d9

remove non-essential requirements (#1103)

n1ck-guo committed 185 days ago

Verified 36749155

move packeding res to cpu (#1104)

yiliu30 committed 185 days ago

Verified daa151d4

refine lmhead ut (#1106)

WeiweiZhang1 committed 186 days ago

Verified 814cf123

Update README.md (#1102)

xin3he committed 186 days ago

Verified e8714e8a

Add Python 3.14 support to compatibility test pipeline (#1096)

XuehaoSun committed 188 days ago

Verified 99bf7720

fix gguf acc and oom bug when iters > 0 (#1098)

n1ck-guo committed 188 days ago

Verified 09d9962a

Refactor input normalization by replaying inputs for consistent preprocessing (#1094)

yiliu30 committed 189 days ago

Verified 152f1a04

add system checker in backend (#1097)

wenhuach21 committed 189 days ago

Verified ae4ac94c

bump version to v0.9.3 (#1091)

chensuyue committed 190 days ago

Verified f81af704

remove accelerate version limitation (#1090)

chensuyue committed 190 days ago

Verified 39b0643b

Fix MXFP/NVFP + FP8 Attn/KV (#1086)

yiliu30 committed 190 days ago

Verified caa246f9

remove numpy restriction for gptq kernel (#1084)

wenhuach21 committed 190 days ago

Verified 0abae385

fix mxfp qdq (#1088)

yiliu30 committed 191 days ago

Verified cd9f20e8

Support alg_ext on windows (#1082)

chensuyue committed 191 days ago

Verified ed07303d

adjust 3 bits hyperparameters at auto-round-best (#1081)

wenhuach21 committed 191 days ago

Verified ef78c688

fix bf16 option in AutoScheme (#1079)

wenhuach21 committed 193 days ago

Verified 5329bcda

refine sampler (#1077)

wenhuach21 committed 195 days ago

Verified 727e573e

Support mxfp nvfp lmhead quant (#1051)

WeiweiZhang1 committed 195 days ago

Verified c4a14799

improve accuracy for 2bit with auto-round-best (#1078)

wenhuach21 committed 196 days ago

Verified 7907c329

Add MXFP8 MOE/Linear and MXFP4 Linear (#1034)

yiliu30 committed 196 days ago

Verified 61bcfdb6

simplify what's new and add publication_list (#1070)

xin3he committed 196 days ago

Verified 7345fe57

update alg_ext and add ut (#1064)

n1ck-guo committed 196 days ago

Verified e18f9b66

Support transformers loading quantized moe model (#1067)

mengniwang95 committed 197 days ago

Verified b1b60e47

Older