Skip to content

Pull requests: sgl-project/sglang

New pull request New
43 Open 1,981 Closed
43 Open 1,981 Closed
Author
Filter by author
Loading
Label
Filter by label
Loading
Use alt + click/return to exclude labels
or ⇧ + click/return for logical OR
Projects
Filter by project
Loading
Milestones
Filter by milestone
Loading
Reviews
Filter by reviews
No reviews Review required Approved review Changes requested
Assignee
Filter by who’s assigned
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

Cutlass w8a8 fp8 kernel high priority
#3047 opened Jan 22, 2025 by HandH1998 Loading…
@ispobock @BBuf
@HandH1998 @zhyncs
12
Fix the FP8 E4M3 parsing offline scales failure bug
#3045 opened Jan 22, 2025 by sleepcoo Loading…
2
Modify the kernel test path & add it to the CI process.
#3044 opened Jan 22, 2025 by sleepcoo Loading…
[Doc]Update doc of profiling with PyTorch Profiler
#3038 opened Jan 22, 2025 by Fridge003 Loading…
2 of 4 tasks
2
Allow local cutlass directory to be used in sgl-kernel build
#3037 opened Jan 22, 2025 by trevor-m Loading…
4 tasks
2
Support int8 kvcahe
#3034 opened Jan 21, 2025 by sleepcoo Loading…
1
More efficient minmax-text-01 lightning_attention_decode with cuda
#3030 opened Jan 21, 2025 by BBuf Loading…
2
Refactor recursive helper methods to iterative approach to prevent s…
#3029 opened Jan 21, 2025 by luzengxiangcn • Draft
4 tasks
Minicpmo
#3023 opened Jan 21, 2025 by mickqian Loading…
3 of 4 tasks
support telechat2 model
#3000 opened Jan 20, 2025 by shunxing12345 Loading…
4 tasks
5
Integrate turbomind into sgl-kernel
#2999 opened Jan 20, 2025 by bjmsong Loading…
4 tasks
@zhyncs
10
[devcontainer] add non-root user
#2989 opened Jan 19, 2025 by ByronHsu • Draft
4 tasks
1
@zhyncs
5
[Fix] Address remain issues of supporting MiniCPMV
#2977 opened Jan 19, 2025 by mickqian Loading…
3 tasks done
4
[MOE] try to optimize cu kernel single block execution - distribute cumsum workload from thread 0 to other threads
#2970 opened Jan 19, 2025 by yiakwy-xpu-ml-framework-team Loading…
3 of 4 tasks
6
[Core] Optimize the delay scheduling of in batch prefix caching
#2962 opened Jan 18, 2025 by MrAta • Draft
4 tasks
Test removing a branch logic
#2905 opened Jan 15, 2025 by rkooo567 Loading…
3 tasks
1
Integration of TurboMind AWQ
#2900 opened Jan 15, 2025 by bjmsong Loading…
3 tasks
1
[Feature] Support dynamic loading and unloading of Lora adapters
#2891 opened Jan 14, 2025 by Fridge003 • Draft
1 of 3 tasks
[DO NOT MERGE] Merged PRs for verl integration
#2849 opened Jan 13, 2025 by fzyzcjy • Draft
3 tasks
Support direct weight loading
#2845 opened Jan 12, 2025 by fzyzcjy Loading…
3 tasks done
Support distributed tensor when updating weights
#2831 opened Jan 10, 2025 by fzyzcjy Loading…
3 tasks done
Support custom device mesh for tensor parallel workers
#2827 opened Jan 10, 2025 by fzyzcjy Loading…
3 tasks done
Use CUDA_VISIBLE_DEVICES instead of gpu_id variables everywhere.
#2824 opened Jan 10, 2025 by heiner Loading…
1 task done
8
Improve the mixed chunk prefill by lanuch two kernels
#2811 opened Jan 9, 2025 by libratiger • Draft
1 of 3 tasks
7
[WIP] [Feature] Support Deepseek-VL2 enhancement New feature or request
#2798 opened Jan 8, 2025 by ccw1996 • Draft
3 tasks
@yizhang2077 @ispobock @zhyncs
4
ProTip! Filter pull requests by the default branch with base:main.