Commits · 18ae428a0d8792d160d811a9cd5bb004d68ea8bd · OpenDAS / vllm_cscc · GitLab

20 Sep, 2024 1 commit
- [Bugfix] Fix Phi3.5 mini and MoE LoRA inference (#8571) · 18ae428a
  Amit Garg authored Sep 19, 2024
  
  18ae428a
19 Sep, 2024 2 commits
- [Core] simplify logits resort in _apply_top_k_top_p (#8619) · e42c634a
  盏一 authored Sep 20, 2024
  
  e42c634a
- Revert "[Misc][Bugfix] Disable guided decoding for mistral tokenizer" (#8593) · 02c9afa2
  Roger Wang authored Sep 18, 2024
  
  02c9afa2
18 Sep, 2024 5 commits
- [Kernel] Change interface to Mamba selective_state_update for continuous batching (#8039) · db9120cd
  Tyler Michael Smith authored Sep 18, 2024
  
  db9120cd
- [AMD][ROCm]Quantization methods on ROCm; Fix _scaled_mm call (#8380) · b3195bc9
  Gregory Shtrasberg authored Sep 18, 2024
```
Co-authored-by: Alexei-V-Ivanov-AMD <156011006+Alexei-V-Ivanov-AMD@users.noreply.github.com>
Co-authored-by: Michael Goin <michael@neuralmagic.com>
```
  b3195bc9
- [Model] Support Solar Model (#8386) · e18749ff
  Geun, Lim authored Sep 19, 2024
```
Co-authored-by: Michael Goin <michael@neuralmagic.com>
```
  e18749ff
- [CI/Build] Update Ruff version (#8469) · 9d104b5b
  Aaron Pham authored Sep 18, 2024
```
Signed-off-by: Aaron Pham <contact@aarnphm.xyz>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
```
  9d104b5b
- [CI/Build] Avoid CUDA initialization (#8534) · 6ffa3f31
  Cyrus Leung authored Sep 18, 2024
  
  6ffa3f31
17 Sep, 2024 6 commits
- [Kernel] Change interface to Mamba causal_conv1d_update for continuous batching (#8012) · 8110e445
  Tyler Michael Smith authored Sep 17, 2024
  
  8110e445
- [Bugfix] Fix TP > 1 for new granite (#8544) · 98f97133
  Joe Runde authored Sep 17, 2024
```
Signed-off-by: Joe Runde <Joseph.Runde@ibm.com>
```
  98f97133
- [Feature][kernel] tensor parallelism with bitsandbytes quantization (#8434) · 9855b995
  chenqianfzh authored Sep 17, 2024
  
  9855b995
- [Encoder decoder] Add cuda graph support during decoding for encoder-decoder models (#7631) · 1009e93c
  sroy745 authored Sep 17, 2024
  
  1009e93c
- [Misc][Bugfix] Disable guided decoding for mistral tokenizer (#8521) · ee2bceaa
  Roger Wang authored Sep 16, 2024
  
  ee2bceaa
- [refactor] remove triton based sampler (#8524) · 546034b4
  Simon Mo authored Sep 16, 2024
  
  546034b4
16 Sep, 2024 2 commits
- [Kernel] AQ AZP 3/4: Asymmetric quantization kernels (#7270) · 5d73ae49
  Luka Govedič authored Sep 16, 2024
  
  5d73ae49
- [Kernel] Enable 8-bit weights in Fused Marlin MoE (#8032) · a091e2da
  ElizaWszola authored Sep 16, 2024
```
Co-authored-by: Dipika <dipikasikka1@gmail.com>
```
  a091e2da
15 Sep, 2024 2 commits
- [Bugfix][Kernel] Add `IQ1_M` quantization implementation to GGUF kernel (#8357) · fc990f97
  Isotr0py authored Sep 16, 2024
  
  fc990f97
- [Bugfix][Model] Fix Python 3.8 compatibility in Pixtral model by updating type annotations (#8490) · 3724d5f6
  Chris authored Sep 15, 2024
  
  3724d5f6
14 Sep, 2024 2 commits
- [torch.compile] add a flag to disable custom op (#8488) · 47790f3e
  youkaichao authored Sep 14, 2024
  
  47790f3e
- [Model] support minicpm3 (#8297) · 8a0cf1dd
  ywfang authored Sep 14, 2024
```
Co-authored-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  8a0cf1dd
13 Sep, 2024 3 commits
- [Hardware][intel GPU] bump up ipex version to 2.3 (#8365) · 85172520
  Kunshang Ji authored Sep 14, 2024
```
Co-authored-by: Yan Ma <yan.ma@intel.com>
```
  85172520
- [Misc] Skip loading extra bias for Qwen2-VL GPTQ-Int8 (#8442) · 06311e29
  Jee Jee Li authored Sep 13, 2024
  
  06311e29
- [BugFix] fix group_topk (#8430) · 8f44a92d
  Dipika Sikka authored Sep 12, 2024
  
  8f44a92d
12 Sep, 2024 7 commits
- [Bugfix] Fix weight loading issue by rename variable. (#8293) · a480939e
  Wenxiang authored Sep 13, 2024
  
  a480939e
- [Hotfix][Pixtral] Fix multiple images bugs (#8415) · d31174a4
  Patrick von Platen authored Sep 13, 2024
  
  d31174a4
- [Hotfix][Core][VLM] Disable chunked prefill by default and prefix caching for... · c1636945
  Roger Wang authored Sep 12, 2024
```
[Hotfix][Core][VLM] Disable chunked prefill by default and prefix caching for multimodal models (#8425)
```
  c1636945
- [Model] Support multiple images for qwen-vl (#8247) · c6202dae
  Alex Brooks authored Sep 12, 2024
```
Signed-off-by: Alex-Brooks <Alex.Brooks@ibm.com>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
Co-authored-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  c6202dae
- [Bugfix] Fix InternVL2 inference with various num_patches (#8375) · e56bf277
  Isotr0py authored Sep 13, 2024
```
Co-authored-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  e56bf277
- [Misc] Use RoPE cache for MRoPE (#8396) · 42ffba11
  Woosuk Kwon authored Sep 11, 2024
  
  42ffba11
- [Gemma2] add bitsandbytes support for Gemma2 (#8338) · 1bf2dd9d
  Blueyo0 authored Sep 12, 2024
  
  1bf2dd9d
11 Sep, 2024 8 commits
- Pixtral (#8377) · d394787e
  Patrick von Platen authored Sep 11, 2024
```
Co-authored-by: Roger Wang <ywang@roblox.com>
```
  d394787e
- [Kernel][Misc] register ops to prevent graph breaks (#6917) · 73202dbe
  bnellnm authored Sep 11, 2024
```
Co-authored-by: Sage Moore <sage@neuralmagic.com>
```
  73202dbe
- [Hardware][Intel] Support compressed-tensor W8A8 for CPU backend (#7257) · 0b952af4
  Li, Jiang authored Sep 12, 2024
  
  0b952af4
- [Model][VLM] Add Qwen2-VL model support (#7905) · 3b7fea77
  Yang Fan authored Sep 12, 2024
```
Co-authored-by: Roger Wang <136131678+ywang96@users.noreply.github.com>
Co-authored-by: DarkLight1337 <tlleungac@connect.ust.hk>
```
  3b7fea77
- [model] Support for Llava-Next-Video model (#7559) · 6a512a00
  Yangshen⚡Deng authored Sep 11, 2024
```
Co-authored-by: Roger Wang <ywang@roblox.com>
Co-authored-by: Cyrus Leung <tlleungac@connect.ust.hk>
Co-authored-by: Cyrus Leung <cyrus.tl.leung@gmail.com>
```
  6a512a00
- [Hardware][NV] Add support for ModelOpt static scaling checkpoints. (#6112) · efcf946a
  Pavani Majety authored Sep 10, 2024
  
  efcf946a
- [Bugfix] Fix InternVL2 vision embeddings process with pipeline parallel (#8299) · 1230263e
  Isotr0py authored Sep 11, 2024
  
  1230263e
- [Misc] Skip loading extra bias for Qwen2-MOE GPTQ models (#8329) · e497b8ae
  Jee Jee Li authored Sep 11, 2024
  
  e497b8ae
10 Sep, 2024 2 commits
- [Bugfix] Fix missing `post_layernorm` in CLIP (#8155) · da1a844e
  Cyrus Leung authored Sep 10, 2024
  
  da1a844e
- [Misc] Fused MoE Marlin support for GPTQ (#8217) · 6cd5e5b0
  Dipika Sikka authored Sep 09, 2024
  
  6cd5e5b0