Commits · ac963be234cb3ff8efc6480f45b45ea08730b89f · change / sglang · GitLab

14 Feb, 2025 4 commits
- update flashinfer-python (#3557) · ac963be2
  Yineng Zhang authored Feb 14, 2025
  
  ac963be2
- chore: bump v0.4.3 (#3556) · e0b9a423
  Yineng Zhang authored Feb 14, 2025
  
  e0b9a423
- chore: bump 0.0.3.post6 sgl-kernel (#3555) · e0821425
  Yineng Zhang authored Feb 14, 2025
  
  e0821425
- feat: support flashinfer mla attention for deepseek v3 (#3550) · 70f894b8
  Yineng Zhang authored Feb 14, 2025
  
  70f894b8
13 Feb, 2025 5 commits
- Update install docs (#3553) · 368de366
  simveit authored Feb 13, 2025
```
Co-authored-by: Chayenne <zhaochen20@outlook.com>
```
  368de366
- update README (#3543) · 20de05a7
  Yineng Zhang authored Feb 13, 2025
  
  20de05a7
- fix moe_align_kernel shm init not sync bug (#3534) · f076328b
  Xiaoyu Zhang authored Feb 13, 2025
  
  f076328b
- Update DeepSeek V3 Doc (#3541) · bf2a7087
  Jhin authored Feb 13, 2025
  
  bf2a7087
- [ROCm] Add ROCm tuning configs for AMD Instinct MI325X. (#3536) · 871a4aa1
  Wen-Heng (Jack) Chung authored Feb 12, 2025
  
  871a4aa1
12 Feb, 2025 9 commits
- integrate blockwise fp8 kernel (#3529) · 98eecbda
  yizhang2077 authored Feb 13, 2025
  
  98eecbda
- chore: bump 0.0.3.post5 sgl-kernel (#3530) · 4430c0a5
  Yineng Zhang authored Feb 13, 2025
  
  4430c0a5
- support blockwise fp8 matmul kernel (#3267) · 640363ad
  yizhang2077 authored Feb 13, 2025
  
  640363ad
- Fix deepseek awq v3 (#3450) · 8616357a
  Liangsheng Yin authored Feb 12, 2025
  
  8616357a
- added llama and cleaned up (#3503) · 8adbc78b
  Zachary Streeter authored Feb 12, 2025
  
  8adbc78b
- use sgl_per_token_group_quant_fp8 kernel (#3493) · 45e3a7bc
  Xiaoyu Zhang authored Feb 12, 2025
  
  45e3a7bc
- chore: bump 0.0.3.post4 sgl-kernel (#3523) · b96e92e6
  Yineng Zhang authored Feb 12, 2025
  
  b96e92e6
- refine deepseek_v3 launch server doc (#3522) · 693c2600
  Xiaoyu Zhang authored Feb 12, 2025
  
  693c2600
- doc: Support a new vLM (#3405) · ced68066
  Mick authored Feb 12, 2025
```
Co-authored-by: zhaochenyang20 <zhaochen20@outlook.com>
```
  ced68066
11 Feb, 2025 12 commits
- Make NCCL NVLS configurable (#3502) · b8318aec
  Ata Fatahi authored Feb 11, 2025
  
  b8318aec
- docs: update install · 2f482210
  Yineng Zhang authored Feb 12, 2025
  
  2f482210
- MI30x: More graph captures for larger batch sizes and concurrencies (#3420) · d81ac443
  HAI authored Feb 11, 2025
  
  d81ac443
- add deepseek-v3 amd docker command (#3495) · 2491cc92
  Zachary Streeter authored Feb 11, 2025
  
  2491cc92
- fix router typo (#3496) · 67c5de92
  Didier Durand authored Feb 11, 2025
  
  67c5de92
- fix server_arguments typo (#3499) · 1e2cf2b5
  Didier Durand authored Feb 11, 2025
  
  1e2cf2b5
- fix supported_models Qwen typo (#3498) · 9490d157
  Didier Durand authored Feb 11, 2025
  
  9490d157
- fix deepseek_v3 typo (#3497) · eefcbdd3
  Didier Durand authored Feb 11, 2025
  
  eefcbdd3
- Support Eagle cuda graph for Triton backend (#3500) · 7e6d5fc6
  Ke Bao authored Feb 12, 2025
  
  7e6d5fc6
- Tune MI300X fused MoE Triton kernel JSON config. (#3492) · cadd5dbe
  Wen-Heng (Jack) Chung authored Feb 11, 2025
  
  cadd5dbe
- optimize per token group quant fp8 (#3490) · bb418ced
  Xiaoyu Zhang authored Feb 11, 2025
  
  bb418ced
- [ROCm] Add ROCm tuning config to block gemm and Re-tune for AMD Radeon Graphics (#3418) · fdf04a14
  yigex authored Feb 11, 2025
```
Co-authored-by: Bruce Xue <yigex@xilinx.com>
Co-authored-by: HAI <hixiao@gmail.com>
```
  fdf04a14
10 Feb, 2025 10 commits
- [Feat] Return hidden states (experimental) (#3364) · 5f0e7de3
  Jackmin801 authored Feb 10, 2025
```
Co-authored-by: Chayenne <zhaochen20@outlook.com>
```
  5f0e7de3
- refine some typo (#3473) · 2f47d710
  Xiaoyu Zhang authored Feb 10, 2025
  
  2f47d710
- fix mla test (#3469) · 4fe92bfc
  Yineng Zhang authored Feb 10, 2025
  
  4fe92bfc
- [Eagle] reduce one draft forward (#3468) · d23cb9a0
  Ying Sheng authored Feb 10, 2025
  
  d23cb9a0
- Support Eagle2 for Triton backend (#3466) · 2d611323
  Ke Bao authored Feb 10, 2025
  
  2d611323
- chore: bump v0.4.2.post4 (#3459) · cddb1cdf
  Yineng Zhang authored Feb 10, 2025
  
  cddb1cdf
- use nvcr.io/nvidia/tritonserver:24.04-py3-min as base image (#3457) · fa1b40e0
  Yineng Zhang authored Feb 10, 2025
  
  fa1b40e0
- [Fix] Fix accuracy bug and refactor codes for lora (#3413) · c45cab1c
  Baizhou Zhang authored Feb 09, 2025
  
  c45cab1c
- remove _grouped_size_compiled_for_decode_kernels (#3453) · 27c4c9cf
  Yineng Zhang authored Feb 10, 2025
  
  27c4c9cf
- Update contribution_guide.md (#3452) · 52a492a1
  Ying Sheng authored Feb 09, 2025
  
  52a492a1