-
Notifications
You must be signed in to change notification settings - Fork 8.1k
Pull requests: sgl-project/sglang
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[CUDA] Add dense and MoE GGUF MMQ kernels for eight I-quant types
sgl-kernel
#36122
opened Aug 24, 2026 by
liuhao-labs
•
Draft
3 of 5 tasks
[Bugfix] Fail fast when hierarchical cache is enabled with FP4 KV cache dtype
#36121
opened Aug 24, 2026 by
jerry-he
Loading…
[NPU] Fix xgrammar apply_vocab_mask device dispatch to use torch.ops.npu
#36120
opened Aug 24, 2026 by
zhujianwei-ops
Loading…
[AMD][DSV4] perf: MXFP8 MoRI dispatch to match the w4a8 MoE input format
#36119
opened Aug 24, 2026 by
karverma-amd
Contributor
•
Draft
4 of 5 tasks
chore: update CI test est_time values
amd
apple-silicon
blackwell
SM100/SM120
deepseek
hicache
Hierarchical Caching for SGLang
lora
memory-pool
Multi-modal
multi-modal language model
npu
quant
LLM Quantization
speculative-decoding
unified-radix-cache
#36117
opened Aug 24, 2026 by
sglang-bot
Member
Loading…
[docs] Slurm: fix multi-node example, add single-node and offline batch examples
documentation
Improvements or additions to documentation
#36116
opened Aug 23, 2026 by
holtwood
Loading…
[AMD][CI] Add a model cache audit that counts per-commit workflows
amd
#36114
opened Aug 23, 2026 by
michaelzhang-ai
Collaborator
•
Draft
4 of 5 tasks
Fix entrypoints/Engine._launch_subprocesses return type annotation
#36113
opened Aug 23, 2026 by
dj707chen
Loading…
2 of 5 tasks
feat(native-mtp): prototype Qwen3-Next native MTP path
#36112
opened Aug 23, 2026 by
ntny
Loading…
5 tasks
[sglang-miles] Add SGLANG_DISABLE_MULTIMEM_AG to force the NCCL all-gather path
#36110
opened Aug 23, 2026 by
yueming-yuan
Collaborator
Loading…
Use model-native video preprocessing for Qwen3-VL
#36109
opened Aug 23, 2026 by
avisingh599
Loading…
[cli] fix: detect model-path from --config in sglang serve
#36107
opened Aug 23, 2026 by
clayrosenthal
Loading…
3 of 5 tasks
[Test] Add unit tests for Triton load watch
#36106
opened Aug 23, 2026 by
chg073
Loading…
5 tasks done
Fix canonical YAML options shadowed by deprecated aliases
#36104
opened Aug 23, 2026 by
gaoxiaomo
Loading…
[perf] Sample next tokens from temperature-scaled logits without materializing full-vocab probs
#36103
opened Aug 23, 2026 by
positive666
Loading…
4 of 5 tasks
[NVIDIA] perf: pipeline the extend-attention KV loads on Hopper
jit-kernel
#36102
opened Aug 23, 2026 by
Hert4
Contributor
Loading…
weight cache: discover daemons by GPU UUID and config
#36101
opened Aug 23, 2026 by
TarangKhanna
•
Draft
[ci] xpu: trigger pr-test-xpu on multimodal_gen changes
run-ci
#36100
opened Aug 23, 2026 by
arathi-hlab
Contributor
Loading…
2 tasks
Add missing test dependencies to pyproject.toml variants
dependencies
Pull requests that update a dependency file
#36098
opened Aug 23, 2026 by
arathi-hlab
Contributor
Loading…
5 tasks
Fix MXFP8 MoE weight sizing for non-gated models
#36097
opened Aug 23, 2026 by
elvischenv
Contributor
Loading…
fix: support MiniMax-M3 pipeline parallelism in PD disaggregation
#36095
opened Aug 23, 2026 by
shiyang814-cpu
Contributor
•
Draft
5 tasks
[AMD][DSV4] perf: retune decode split-K heuristic for MI355X
jit-kernel
#36094
opened Aug 23, 2026 by
karverma-amd
Contributor
Loading…
5 tasks done
Previous Next
ProTip!
Filter pull requests by the default branch with base:main.