Commits · 79b6fbd84aa5a06b0e9ea3a829fc116444ff64b9 · OpenDAS / AutoAWQ

16 Feb, 2024 4 commits
- Support Fused Mixtral on multi-GPU (#352) · 79b6fbd8
  Casper authored Feb 16, 2024
  
  79b6fbd8
- Bump to 0.2.1 (#351) · 74053104
  Casper authored Feb 16, 2024
  
  74053104
- Fix triton dependency (#350) · 8849043a
  Casper authored Feb 16, 2024
  
  8849043a
- ENH / FIX: Few enhancements and fix for mixed-precision training (#348) · 969b290e
  Younes Belkada authored Feb 16, 2024
  
  969b290e
15 Feb, 2024 7 commits
- Avoid downloading ROCm (#347) · 2de6092a
  Casper authored Feb 15, 2024
  
  2de6092a
- Fix typo in setup.py (#346) · d8311027
  Casper authored Feb 15, 2024
  
  d8311027
- Fix workflow (#345) · be0f9e01
  Casper authored Feb 15, 2024
  
  be0f9e01
- Fix importlib (#344) · 171c7aff
  Casper authored Feb 15, 2024
  
  171c7aff
- Fix dependency (#343) · 890127fa
  Casper authored Feb 15, 2024
  
  890127fa
- Fix ROCm build (#342) · 9477ea5c
  Casper authored Feb 15, 2024
  
  9477ea5c
- v0.2.0 (#330) · bcaa8a36
  Casper authored Feb 15, 2024
```
Co-authored-by: jinz2014 <7799920+jinz2014@users.noreply.github.com>
Co-authored-by: Jin Z <5zj@cousteau.ftpn.ornl.gov>
```
  bcaa8a36
14 Feb, 2024 1 commit
- [`PEFT`] Fix PEFT batch size > 1 (#338) · c69d3b65
  Younes Belkada authored Feb 14, 2024
  
  c69d3b65
03 Feb, 2024 5 commits
- PEFT compatible GEMM (#324) · 29ee66d9
  Casper authored Feb 03, 2024
  
  29ee66d9
- Fix n_samples (#326) · ebe8fc3f
  Casper authored Feb 03, 2024
  
  ebe8fc3f
- Add qwen2 (#321) · cc75d0e8
  Junyang Lin authored Feb 03, 2024
  
  cc75d0e8
- Marlin symmetric quantization and inference (#320) · 34085edc
  Ilyas Moutawwakil authored Feb 03, 2024
```
Co-authored-by: Casper <casperbh.96@gmail.com>
```
  34085edc
- AMD ROCM Support (#315) · f018d2b7
  Ilyas Moutawwakil authored Feb 03, 2024
  
  f018d2b7
26 Jan, 2024 1 commit
- Fixes a breaking change in autoawq (#325) · 8c78db4d
  Younes Belkada authored Jan 26, 2024
  
  8c78db4d
24 Jan, 2024 3 commits
- Evaluation: Add more evals (#283) · 7133575a
  Casper authored Jan 24, 2024
  
  7133575a
- Up to 60% faster context processing (#316) · 149236e4
  Casper authored Jan 24, 2024
  
  149236e4
- Torch only inference + any-device quantization (#319) · c6c7b065
  Casper authored Jan 24, 2024
  
  c6c7b065
21 Jan, 2024 2 commits
- Cleanup requirements (#295) · 8117845b
  Casper authored Jan 21, 2024
  
  8117845b
- Exllama kernels support (#313) · 2fcbf268
  Ilyas Moutawwakil authored Jan 21, 2024
```
Co-authored-by: Casper <casperbh.96@gmail.com>
```
  2fcbf268
07 Jan, 2024 1 commit
- GGUF compatible quantization (2, 3, 4 bit / any bit) (#285) · a3db8099
  Casper authored Jan 07, 2024
  
  a3db8099
04 Jan, 2024 1 commit
- Add CPU-loaded multi-GPU quantization (#289) · 46415f5a
  Blake Wyatt authored Jan 04, 2024
  
  46415f5a
28 Dec, 2023 1 commit
- AWQ: Separate the AWQ kernels to separate repository (#279) · 2edb3f6f
  Casper authored Dec 28, 2023
```
Co-authored-by: Casper Hansen <casperbh96@gmail.com>
```
  2edb3f6f
23 Dec, 2023 6 commits
- Bump to v0.1.8 (#274) · 3f10cf1d
  Casper authored Dec 23, 2023
  
  3f10cf1d
- Add vLLM async example (#273) · de166044
  Casper authored Dec 23, 2023
  
  de166044
- Update news and models supported (#272) · 78577f54
  Casper authored Dec 23, 2023
  
  78577f54
- Set default rope_theta on LlamaLikeBlock (#271) · 66adba4a
  Casper authored Dec 23, 2023
  
  66adba4a
- Add Baichuan2 Support (#247) · cef9f113
  Aoyu authored Dec 23, 2023
```
Co-authored-by: Casper <casperbh.96@gmail.com>
```
  cef9f113
- FEAT: add llava to autoawq (#250) · 9e8e28b2
  Younes Belkada authored Dec 23, 2023
```
Co-authored-by: Casper <casperbh.96@gmail.com>
```
  9e8e28b2
22 Dec, 2023 2 commits
- Fused rope theta (#270) · 727172e9
  Casper authored Dec 23, 2023
```
Co-authored-by: Casper Hansen <casperbh96@gmail.com>
```
  727172e9
- Mixtral: Mixture of Experts quantization (#251) · 5b9f3c47
  Casper authored Dec 22, 2023
  
  5b9f3c47
16 Dec, 2023 1 commit
- Fix quantization issue with transformers >= 4.36.0 (#264) · 2350a4d0
  Younes Belkada authored Dec 16, 2023
  
  2350a4d0
11 Dec, 2023 3 commits
- FEAT: Add possibility of skipping modules when quantizing (#248) · 9c3dfa07
  Younes Belkada authored Dec 11, 2023
  
  9c3dfa07
- [`core`] Fix quantization issues with transformers==4.36.0 (#249) · 78b59d73
  Younes Belkada authored Dec 11, 2023
  
  78b59d73
- [`core`] Make AutoAWQ fused modules compatible with HF transformers (#244) · ae24f424
  Younes Belkada authored Dec 11, 2023
  
  ae24f424
08 Dec, 2023 2 commits
- Update README.md (#245) · 09db054f
  Casper authored Dec 08, 2023
  
  09db054f
- Load on CPU to avoid OOM (#236) · 6b5dc29f
  Casper authored Dec 08, 2023
  
  6b5dc29f