Commits · cddb1cdf8fd85538003990dd12ab4f686d3da064 · zhaoyu6 / sglang

"vscode:/vscode.git/clone" did not exist on "3aec3d4f8b9f4c2cc84df13b71c9880a4ec2a6aa"

10 Feb, 2025 3 commits
- chore: bump v0.4.2.post4 (#3459) · cddb1cdf
  Yineng Zhang authored Feb 10, 2025
  
  cddb1cdf
- [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
09 Feb, 2025 8 commits
- feat: enable ragged fa3 by default on hopper 12.4+ (#3442) · 36f6fc50
  Yineng Zhang authored Feb 10, 2025
  
  36f6fc50
- support version in sgl-kernel (#3439) · 4cfd3add
  Yineng Zhang authored Feb 10, 2025
  
  4cfd3add
- compatible with new outlines (#3435) · 85986bb9
  Yineng Zhang authored Feb 10, 2025
  
  85986bb9
- remove activation dependency in fused_moe (#3433) · 64c87135
  Yineng Zhang authored Feb 10, 2025
  
  64c87135
- fix draft cuda graph capture failure (#3431) · 1646149a
  Yineng Zhang authored Feb 09, 2025
  
  1646149a
- add cuda graph capture failure possible solution (#3430) · bc72e5bd
  Yineng Zhang authored Feb 09, 2025
  
  bc72e5bd
- update forward_return_lse (#3425) · 014cab4d
  Yineng Zhang authored Feb 09, 2025
  
  014cab4d
- [Fix] Fix eagle with disable cuda graph (#3411) · 7b4e61ff
  Ying Sheng authored Feb 08, 2025
  
  7b4e61ff
08 Feb, 2025 2 commits
- fix EAGLE 2 non greedy case (#3407) · fad315cb
  Yineng Zhang authored Feb 09, 2025
```
Co-authored-by: Ying Sheng <sqy1415@gmail.com>
```
  fad315cb
- Add H20 fp8 w8a8 gemm config (#3386) · 4530136e
  lukec authored Feb 08, 2025
  
  4530136e
07 Feb, 2025 7 commits
- chore: bump v0.4.2.post3 (#3369) · c1f5f99f
  Yineng Zhang authored Feb 08, 2025
  
  c1f5f99f
- fix EagleVerifyInput (#3378) · fa82dfcc
  Yineng Zhang authored Feb 07, 2025
  
  fa82dfcc
- update sgl-kernel version (#3374) · f2870376
  Yineng Zhang authored Feb 07, 2025
  
  f2870376
- support speculative decoding kernel in sgl-kernel (#3373) · f9905d59
  Yineng Zhang authored Feb 07, 2025
```
Co-authored-by: Ying Sheng <sqy1415@gmail.com>
```
  f9905d59
- update waves_per_eu to 1 (#3356) · e868d0b6
  lizamd authored Feb 06, 2025
  
  e868d0b6
- Fix: Runtime error for function calling (#3300) · 591e751e
  Shi Shuai authored Feb 07, 2025
  
  591e751e
- Feature: Fix the binding error in Llama (#3355) · 40022d07
  Chayenne authored Feb 06, 2025
  
  40022d07
06 Feb, 2025 2 commits
- optimize moe_align_kernel cuda (#3347) · cdae77b0
  Xiaoyu Zhang authored Feb 07, 2025
  
  cdae77b0
- [ROCm] Fix fp8 unrolledx4 matmul kernel. (#3325) · 32de54ed
  Wen-Heng (Jack) Chung authored Feb 05, 2025
```
Co-authored-by: HAI <hixiao@gmail.com>
```
  32de54ed
05 Feb, 2025 6 commits
- Support custom mask for Triton attention (#3317) · a322051e
  Ke Bao authored Feb 06, 2025
  
  a322051e
- Update Triton extend backend interface (#3309) · de553334
  Ke Bao authored Feb 05, 2025
  
  de553334
- chore: bump v0.4.2.post2 (#3313) · 7aad8d18
  Yineng Zhang authored Feb 05, 2025
  
  7aad8d18
- Fix lora flashinfer import bug on ROCM (#3312) · 76fa2d15
  Baizhou Zhang authored Feb 05, 2025
  
  76fa2d15
- [ROCm] Logic to decide whether to used manually unrolled kernel. (#3306) · 7ab84948
  Wen-Heng (Jack) Chung authored Feb 04, 2025
  
  7ab84948
- Use forward_cuda to execute custom op for hip platform (#3305) · 4885b908
  kk authored Feb 05, 2025
```
Co-authored-by: wunhuang <wunhuang@amd.com>
```
  4885b908
04 Feb, 2025 6 commits
- [ROCm] Manually unroll _w8a8_block_fp8_matmul kernel on AMD GPU. (#3299) · c2723a42
  Wen-Heng (Jack) Chung authored Feb 04, 2025
  
  c2723a42
- [ROCm] Add tuning configs for AMD Radeon Graphics. (#3294) · c7256ca8
  Wen-Heng (Jack) Chung authored Feb 04, 2025
  
  c7256ca8
- Update Triton decode backend interface (#3292) · a07364cc
  Ke Bao authored Feb 04, 2025
  
  a07364cc
- ROCm: sgl-kernel enablement starting with sgl_moe_align_block (#3287) · 2c1a695f
  HAI authored Feb 04, 2025
  
  2c1a695f
- upgrade flashinfer v0.2.0.post2 (#3288) · d39899e8
  Yineng Zhang authored Feb 04, 2025
```
Co-authored-by: pankajroark <pankajroark@users.noreply.github.com>
```
  d39899e8
- [Feature] Define backends and add Triton backend for Lora (#3161) · 70817a7e
  Baizhou Zhang authored Feb 03, 2025
```
Co-authored-by: Ying Sheng <sqy1415@gmail.com>
```
  70817a7e
03 Feb, 2025 2 commits
- adding Triton configs for DeepSeekV3 on Blackwell (#3272) · d54cee14
  kushanam authored Feb 03, 2025
  
  d54cee14
- refactor EAGLE 2 (#3269) · 013021b6
  Yineng Zhang authored Feb 03, 2025
```
Co-authored-by: Ying Sheng <sqy1415@gmail.com>
Co-authored-by: merrymercy <lianminzheng@gmail.com>
Co-authored-by: Ying1123 <sqy1415@gmail.com>
```
  013021b6
02 Feb, 2025 3 commits
- Bug: Fix min_p sampling crash when using flashinfer backend (#3207) · 28b0a62b
  zifeitong authored Feb 02, 2025
```
Co-authored-by: zhaochenyang20 <zhaochen20@outlook.com>
```
  28b0a62b
- ROCm: bump 6.3.0 (#3259) · 566d61d9
  HAI authored Feb 02, 2025
  
  566d61d9
- Tune paged attention parameters for AMD GPU. (#3255) · d9eb9358
  Wen-Heng (Jack) Chung authored Feb 01, 2025
  
  d9eb9358
01 Feb, 2025 1 commit
- use srt VocabParallelEmbedding (#3252) · 959dca4f
  Yineng Zhang authored Feb 01, 2025
  
  959dca4f