intel/auto-round

Pull Requests Commits

Fix exlllamav2 backend issue (#144)

wenhuach21 committed 2 years ago

Verified e614d138

Fix asym kernel issue by following autogptq's pr (#137)

wenhuach21 committed 2 years ago

Verified 794cd903

fix typos (#140)

WeiweiZhang1 committed 2 years ago

Verified 4d2d2591

bump version into v0.2 (#139)

chensuyue committed 2 years ago

Verified aafb82ef

handling transformers version compatibility in lmhead export, bugfix (#130)

WeiweiZhang1 committed 2 years ago

Verified 4db22e1d

fix export issue with torch 2.0 (#129)

wenhuach21 committed 2 years ago

Verified 5bff86ee

Update falcon recipe (#128)

wenhuach21 committed 2 years ago

Verified 416ec7e9

fix falcon quant issue with disable_trust_remote_code (#126)

WeiweiZhang1 committed 2 years ago

Verified e2985fdf

Update phi2 recipe (#124)

wenhuach21 committed 2 years ago

Verified edca2980

remove fp32 conversion in exporting to autogptq (#123)

wenhuach21 committed 2 years ago

Verified 17024b16

update gemma recipe (#121)

wenhuach21 committed 2 years ago

Verified ecc1dd65

Remove unused hook (#122)

XuehaoSun committed 2 years ago

Verified c7751c49

support `transformers.Conv1D` packing (#118)

Kaihui-intel committed 2 years ago

Verified 02a6660e

Fix export format issue (#120)

wenhuach21 committed 2 years ago

Verified ed29cf50

wenhuach21 committed 2 years ago

Verified 7bc9d8fb

fix lm-head quant issue at disable_quanted_input (#117)

wenhuach21 committed 2 years ago

Verified b51cfa98

support real lm-head quantization and mixed precision inference (#114)

wenhuach21 committed 2 years ago

Verified 4d1caebb

fix lm-head gradient accumulation bug (#113)

wenhuach21 committed 2 years ago

Verified ecca5349

update shells (#112)

WeiweiZhang1 committed 2 years ago

Verified 3c214db0

Adjust the default evaluation data type by selecting from the model path configuration (#107)

WeiweiZhang1 committed 2 years ago

Verified c42eaa92

20% speedup by removing new zero tensor (#110)

wenhuach21 committed 2 years ago

Verified c7434b63

1.8X speedup by disable_low_gpu_mem_usage and reduce memory usage by avoid using torch.cat (#106)

wenhuach21 committed 2 years ago

Verified 23b60c3a

remove costly operations

wenhuach21 committed 2 years ago

Verified ad3a7bba

Consolidate dataloader&dataset_split to dataset (#105)

wenhuach21 committed 2 years ago

Verified 1f9cb4f8

disable quantizing lm-head with tied weights as a workaround (#102)

wenhuach21 committed 2 years ago

Verified 1fe6aae5

disable quantizing lm-head with tied weights as a workaround (#101)

wenhuach21 committed 2 years ago

Verified c839e825

update readme of calibration dataset and lm-head usage (#98)

wenhuach21 committed 2 years ago

Verified f95b8c7d

fix critic bug for gradient_accumulate_steps!=1 and reduce cpu memory of lm-head tuning (#97)

WeiweiZhang1 committed 2 years ago

Verified 7dd02eb4

handle invalid layername in weight_config (#93)

WeiweiZhang1 committed 2 years ago

Verified 89562261

yintong-lu committed 2 years ago

Verified 9b08de48

Older