-
Notifications
You must be signed in to change notification settings - Fork 5k
Pull requests: sgl-project/sglang
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
test: add unit tests for managers/utils.py and template_manager.py
#21227
opened Mar 23, 2026 by
Xio-Shark
Loading…
chore: migrate callers from /get_server_info to /server_info
amd
deepseek
documentation
Improvements or additions to documentation
hicache
Hierarchical Caching for SGLang
model-gateway
speculative-decoding
#21226
opened Mar 23, 2026 by
Xio-Shark
Loading…
[Spec][Ngram] 4/N: Remove Improvements or additions to documentation
lora
speculative-decoding
max_match_window_size and min_match_window_size, matching all suffixes of the Trie
documentation
#21225
opened Mar 23, 2026 by
kpham-sgl
Loading…
5 tasks done
fix: abort over-length embedding and grammar errors without crash
#21224
opened Mar 23, 2026 by
yang1002378395-cmyk
Loading…
fix: prevent KeyError in _handle_abort_req when request already completed
#21223
opened Mar 23, 2026 by
yang1002378395-cmyk
Loading…
feat: support robust ModelExpress metadata API for RDMA transfer
#21222
opened Mar 23, 2026 by
AndyDai-nv
•
Draft
5 tasks
fix: skip missing expert params in PP mode for Qwen3.5 MoE
#21220
opened Mar 23, 2026 by
yang1002378395-cmyk
Loading…
4 tasks done
[Fix] nvcc compile error in c++17
jit-kernel
#21216
opened Mar 23, 2026 by
Capronir
Loading…
5 tasks
fix: disable attn_tp_input_scattered when input_embeds is provided externally for Kimi-K2.5
deepseek
#21215
opened Mar 23, 2026 by
qingchanghan
Loading…
3 of 5 tasks
[Diffusion][Feat] support LoKr
diffusion
SGLang Diffusion
lora
#21214
opened Mar 23, 2026 by
RuixiangMa
Loading…
5 tasks
[AMD]: Support MLA with nhead<16 and FP8 KV cache for TP=8 (Kimi K2.5…
#21213
opened Mar 23, 2026 by
ZiguanWang
Loading…
5 tasks done
[Bugfix][NPU] Skip FRACTAL_NZ format for MoE weights with unaligned dimensions
quant
LLM Quantization
#21209
opened Mar 23, 2026 by
adityavaid
Loading…
2 of 5 tasks
[Bugfix] restore EAGLE draft padding state between speculative decode steps
#21207
opened Mar 23, 2026 by
lviy
Loading…
3 of 5 tasks
[RaidxTree Refactor]: Support Unified HybridRadixTree V2
high priority
run-ci
#21206
opened Mar 23, 2026 by
hzh0425
Loading…
5 tasks
[Diffusion] Revamp Rollout Log-Prob Support with SDE/CPS for RL Post-Training
diffusion
SGLang Diffusion
#21204
opened Mar 23, 2026 by
Rockdu
Loading…
[KDA] Support CuTeDSL KDA decode kernel
jit-kernel
run-ci
#21203
opened Mar 23, 2026 by
yuan-luo
Loading…
5 tasks
style refinement for hisparse
jit-kernel
run-ci
#21198
opened Mar 23, 2026 by
xiezhq-hermann
Loading…
5 tasks
[NPU]adaptation to support deterministic inference
deterministic
Issues on deterministic inference/kernels
npu
#21197
opened Mar 23, 2026 by
Estrella-xx
•
Draft
1 of 5 tasks
[bugfix][AMD] Fix PPMissingLayer AttributeError for deepseek v2/v3 in aiter_gfx95 code path
deepseek
#21194
opened Mar 23, 2026 by
RolaoDenthu
Loading…
5 tasks
[AMD] Fix AMD Nightly Test - Transformers 5.3.0 incompatibility and gemma2-27b kv issue
#21193
opened Mar 23, 2026 by
yctseng0211
•
Draft
5 tasks
fix: Use base GPU ID CUDA device for multimodal processor
#21191
opened Mar 23, 2026 by
moehanabi
Loading…
2 of 5 tasks
[Whisper] Enable CUDA graph support for encoder-decoder models
blackwell
SM100/SM120
#21190
opened Mar 23, 2026 by
JustinTong0323
Loading…
4 tasks done
Previous Next
ProTip!
no:milestone will show everything without a milestone.