Commits · 36e35facd6720b31c10b3e5bf14222bf25cd8744 · OpenDAS / vllm_cscc · GitLab

11 Feb, 2025 5 commits
- 修改错误代码 · 36e35fac
  王敏 authored Feb 11, 2025
  
  36e35fac
- 添加ep moe tuning · 10213be4
  王敏 authored Feb 11, 2025
  
  10213be4
- [feat]添加ep moe功能 · d0dafaf4
  王敏 authored Feb 11, 2025
  
  d0dafaf4
- update fused_moe.py · a27fdb55
  zhuwenwen authored Feb 11, 2025
  
  a27fdb55
- add latency · 49bfe4cb
  zhuwenwen authored Feb 11, 2025
  
  49bfe4cb
10 Feb, 2025 2 commits
- [feat]适配0.7.2版本的deepseek_v3 nn layout · 70deb5ef
  王敏 authored Feb 10, 2025
  
  70deb5ef
- update model layout · 99d49945
  zhuwenwen authored Feb 10, 2025
  
  99d49945
09 Feb, 2025 1 commit
- update llama and qwen2 layout · 92d43fd5
  zhuwenwen authored Feb 09, 2025
  
  92d43fd5
08 Feb, 2025 4 commits
- set TRITON_HIP_USE_NEW_STREAM_PIPELINE · 8f8d56c3
  zhuwenwen authored Feb 08, 2025
  
  8f8d56c3
- update layer.py · 54944679
  zhuwenwen authored Feb 08, 2025
  
  54944679
- [feat] add nn_moe · 37b63c24
  zhuwenwen authored Feb 08, 2025
  
  37b63c24
- add bw fused_moe_config · 2dc7ec2f
  zhuwenwen authored Feb 08, 2025
  
  2dc7ec2f
06 Feb, 2025 10 commits
- update chatglm nn layout · caf953b6
  zhuwenwen authored Feb 06, 2025
  
  caf953b6
- [Misc] Improve error message for incorrect pynvml (#12809) · 0408efc6
  youkaichao authored Feb 06, 2025
```
Signed-off-by: youkaichao <youkaichao@gmail.com>
```
  0408efc6
- skip arch reason · 66a7ebd8
  zhuwenwen authored Feb 06, 2025
  
  66a7ebd8
- [Misc] Remove duplicated DeepSeek V2/V3 model definition (#12793) · 449d1bce
  Michael Goin authored Feb 06, 2025
  
  449d1bce
- Improve `TransformersModel` UX (#12785) · 1a6fcad4
  Harry Mellor authored Feb 06, 2025
  
  1a6fcad4
- [Model] LoRA Support for Ultravox model (#11253) · d88506dd
  Sumit Vij authored Feb 05, 2025
  
  d88506dd
- [Misc][Easy] Remove the space from the file name · 9cdea30b
  Lu Fang authored Feb 05, 2025
  
  9cdea30b
- [Bugfix] Better FP8 supported defaults · 76abd0c8
  Lucas Wilkinson authored Feb 05, 2025
  
  76abd0c8
- [VLM] Update compatibility with transformers 4.49 · 75404d04
  Cyrus Leung authored Feb 06, 2025
  
  75404d04
- update rocm_flash_attn.py · 49be5e62
  zhuwenwen authored Feb 06, 2025
  
  49be5e62
05 Feb, 2025 16 commits
- [VLM] Qwen2.5-VL · bf3b79ef
  Roger Wang authored Feb 05, 2025
  
  bf3b79ef
- [VLM] Use shared field to pass token ids to model · a4ce74c1
  Cyrus Leung authored Feb 06, 2025
  
  a4ce74c1
- Add: Support for Sparse24Bitmask Compressed Models · 3b2005e1
  Rahul Tuli authored Feb 05, 2025
  
  3b2005e1
- [Hardware][Intel-Gaudi] Enable FusedSDPA support for Intel Gaudi (HPU) · af8486de
  Sanju C Sudhakaran authored Feb 06, 2025
  
  af8486de
- Merging PR #12536 · 4c3aac51
  Chen Zhang authored Feb 06, 2025
```
Merged via CLI script
```
  4c3aac51
- [core][distributed] exact ray placement control (#12732) · bc1bdece
  youkaichao authored Feb 06, 2025
```
Signed-off-by: youkaichao <youkaichao@gmail.com>
```
  bc1bdece
- [Bugfix] Fix 'ModuleNotFoundError: No module named... · 022bcc70
  Akash kaothalkar authored Feb 05, 2025
```
[Bugfix] Fix 'ModuleNotFoundError: No module named 'intel_extension_for_pytorch'' for --tensor-parallel-size more than 1  (#12546)
```
  022bcc70
- [V1][Misc] Shorten `FinishReason` enum and use constant strings (#12760) · 3d09e592
  Nick Hill authored Feb 04, 2025
  
  3d09e592
- [Bugfix] Fix OpenVINO model runner (#12750) · fcf2e3d7
  Harry Mellor authored Feb 05, 2025
  
  fcf2e3d7
- [Model][Quant] Fix GLM, Fix fused module mappings for quantization (#12634) · 7ff7a638
  Kyle Sayers authored Feb 05, 2025
```
Signed-off-by: mgoin <michael@neuralmagic.com>
Signed-off-by: Kyle Sayers <kylesayrs@gmail.com>
Co-authored-by: mgoin <michael@neuralmagic.com>
```
  7ff7a638
- [VLM] Add MLA with pure RoPE support for deepseek-vl2 models (#12729) · 98fd089f
  Isotr0py authored Feb 05, 2025
  
  98fd089f
- Refactor `Linear` handling in `TransformersModel` (#12727) · 249824c3
  Harry Mellor authored Feb 05, 2025
```
Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
```
  249824c3
- [ROCM][AMD][TRITON] Halving warps number for fw_prefill to reduce spilling (#12713) · 64862d10
  Aleksandr Malyshev authored Feb 04, 2025
```
Signed-off-by: Aleksandr Malyshev <maleksan@amd.com>
Co-authored-by: Aleksandr Malyshev <maleksan@amd.com>
```
  64862d10
- [Core] add and implement `VLLM_LOGITS_PROCESSOR_THREADS` (#12368) · b3a0d01e
  Aviv Keshet authored Feb 04, 2025
```
Signed-off-by: Aviv Keshet <akeshet@scaledcognition.com>
```
  b3a0d01e
- [Perf] Mem align KV caches for CUDA devices (MLA perf improvement) (#12676) · 75e94309
  Lucas Wilkinson authored Feb 04, 2025
```
Signed-off-by: simon-mo <xmo@berkeley.edu>
Signed-off-by: Lucas Wilkinson <lcwilkins@redhat.com>
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Signed-off-by: Lucas Wilkinson <lwilkinson@neuralmagic.com>
Co-authored-by: simon-mo <xmo@berkeley.edu>
```
  75e94309
- [V1][Metrics] Add request_success_total counter, labelled with finish reason (#12579) · 233df6f5
  Mark McLoughlin authored Feb 05, 2025
```
Signed-off-by: Mark McLoughlin <markmc@redhat.com>
```
  233df6f5
04 Feb, 2025 2 commits
- Avoid unnecessary multi-modal input data copy when len(batch) == 1 (#12722) · 62467a83
  Kero Liang authored Feb 04, 2025
```
Signed-off-by: imkero <kerorek@outlook.com>
```
  62467a83
- [Bugfix] Fix loading of fine-tuned models based on Phi-3-Small (#12689) · 6469038b
  Michael Greenbaum authored Feb 04, 2025
```
Signed-off-by: Michael Greenbaum <mgreenbaum@microsoft.com>
Co-authored-by: Michael Greenbaum <mgreenbaum@microsoft.com>
```
  6469038b