Pull Requests casper-hansen/AutoAWQ

the final piece (v2)

#761 by casper-hansen was merged 2025-05-11 08:52

the final piece

#760 by casper-hansen was merged 2025-05-11 08:29

Add Qwen2.5-Omni support.

#759 by tiger-of-shawn was merged 2025-05-11 07:11

Update quantizer.py for saving vram while quantization

#757 by kIshizaki-sci was closed 2025-05-11 09:10

update scripts

#756 by casper-hansen was merged 2025-05-03 09:41

assertion for non-activated experts in MoE

#755 by casper-hansen was merged 2025-05-03 09:29

Added Qwen 3 support.

#751 by zju-stu-lizheng was merged 2025-04-29 06:43

NEWS

#750 by casper-hansen was merged 2025-04-24 11:33

Add llama4 support

#748 by kIshizaki-sci was closed 2025-05-04 08:12

fix: mrope cause shpae is not match

#746 by ppppppppig was closed 2025-05-11 09:10

Add Jupyter Notebook support for rich tqdm progress bar display

#744 by kIshizaki-sci was closed 2025-05-04 08:27

add internlm3 support

#743 by BBC-Esq was closed 2025-04-24 20:58

fix: mrope cause shpae is not match

#742 by ppppppppig was closed 2025-04-12 13:39

fix dtype mismatch

#740 by jiqing-feng was merged 2025-04-09 08:42

Twostep quant

#720 by oj9040 was closed 2025-02-26 05:17

fix(cache): remove k dim in cache

#718 by neurowelt was merged 2025-03-06 15:13

fix setup

#715 by jiqing-feng was merged 2025-03-06 15:11

Fixed issue with the slow implementation warning

#711 by Egor-Krivov was merged 2025-03-06 15:11

Add Qwen2.5-VL

#706 by seungwoos was merged 2025-03-06 15:10

Add computed position embedding external

#705 by seungwoos was closed 2025-02-07 14:22

bump to 028

#700 by casper-hansen was merged 2025-01-20 10:43

fix workflow build

#699 by casper-hansen was merged 2025-01-20 10:15

automatically load dtype from config

#698 by casper-hansen was merged 2025-01-20 08:47

add ability to define torch_dtype

#697 by casper-hansen was merged 2025-01-20 08:21

fix bug when using FSDP

#696 by kaixuanliu was merged 2025-01-20 06:42

Enable triton on XPU devices

#695 by Egor-Krivov was merged 2025-01-20 08:51

add support for telechat2

#694 by 1096125073 was closed 2025-05-11 09:10

Added DeepSeek V3 support.

#688 by LagPixelLOL was merged 2025-01-20 08:07

[WIP] Pixtral

#681 by casper-hansen was closed 2025-05-11 09:10

improve type hinting and fix use_cache

#680 by casper-hansen was merged 2024-12-13 21:17