Skip to content

Pull requests: sgl-project/sglang-jax

New pull request New
46 Open 586 Closed
46 Open 586 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
Assigned to nobody Loading
Sort
Sort by
Newest Oldest Most commented Least commented Recently updated Least recently updated Best match
Most reactions
👍 👎 😄 🎉 😕 ❤️ 🚀 👀

Pull requests list

deepseekv3 update
#1019 opened May 3, 2026 by Bob-Chen222 Collaborator Loading…
4 tasks
@Bob-Chen222
4
perf(moe): align expert tokens to gmm tile size once upstream
#1012 opened Apr 30, 2026 by alexnails Collaborator Loading…
4 tasks
7
fix(model_runner): fail-fast when hybrid recurrent has 0 full-attention layers
#1005 opened Apr 29, 2026 by Rodrian7 Collaborator • Draft
3 tasks done
Adding RPAv3 tuner
#987 opened Apr 27, 2026 by depksingh Contributor Loading…
4 tasks done
1
5
[WIP] Epic/support kimi linear
#985 opened Apr 27, 2026 by aolemila Collaborator Loading…
4 tasks
1
refactor(quant): simplify imports and rename channelwise kernel
#977 opened Apr 27, 2026 by cjx0709 Collaborator Loading…
3 tasks
2
1
add kda prefill and decode
#964 opened Apr 25, 2026 by pathfinder-pf Collaborator Loading…
4 tasks
1
Feat/rm dequant
#960 opened Apr 24, 2026 by pengchengneo Collaborator Loading…
4 tasks
1
feat: Allow sequence parallel and enable it for grok-2
#952 opened Apr 23, 2026 by hlFu Collaborator Loading…
3 of 4 tasks
9
fix: linear attention bug fixes (head_dim, rotary_dim, slope, T_packed_bucket)
#928 opened Apr 20, 2026 by Rodrian7 Collaborator Loading…
4 tasks
1
1
[draft] bailing-v2.5
#925 opened Apr 17, 2026 by Fred33146 Loading…
4 tasks
1
fix: remove process_count guards for multi-process sharding
#919 opened Apr 16, 2026 by JamesBrianD Collaborator Loading…
3 tasks
2
Deepseekv3
#916 opened Apr 16, 2026 by Bob-Chen222 Collaborator • Draft
4 tasks
@Bob-Chen222
1
feat: Add ngram speculative decoding (Prompt Lookup Decoding, no draft model)
#908 opened Apr 12, 2026 by hashkanna Loading…
3 of 4 tasks
3
fix: raise NotImplementedError for attention_sink in fused KV path + add CPU correctness tests
#906 opened Apr 10, 2026 by zorrofox Contributor Loading…
1 of 2 tasks
3
Support Flux 1.dev
#902 opened Apr 10, 2026 by bjmsong Contributor Loading…
6
wip: test selective CI testing
#890 opened Apr 3, 2026 by zhengkezhou1 Collaborator Loading…
4 tasks
1
Feat/dtype config
#888 opened Apr 2, 2026 by iwknow Collaborator Loading…
3 tasks done
15
attention optimization by upgrading libtpu and jax
#878 opened Mar 18, 2026 by Bob-Chen222 Collaborator Loading…
4 tasks
3
feat: support per-layer hidden states extraction via Engine API
#877 opened Mar 18, 2026 by zhengkezhou1 Collaborator Loading…
4 tasks
1
2
docs: add guide for running TPU tests natively through AI Agent
#872 opened Mar 13, 2026 by zhengkezhou1 Collaborator • Draft
4 tasks
[WIP] update parameters in get_block_spec_config.py
#864 opened Mar 5, 2026 by aolemila Collaborator Loading…
4 tasks
1
1
[WIP] enable simple tune and change field initialization position
#863 opened Mar 5, 2026 by aolemila Collaborator Loading…
4 tasks
1
1
quantized matmul kernel integration
#841 opened Feb 26, 2026 by Bob-Chen222 Collaborator • Draft
4 tasks
1
feat: Add support for LTX-2 Video Diffusion Model
#840 opened Feb 26, 2026 by chandrasekhard2 Loading…
6 tasks done
4
ProTip! Exclude everything labeled bug with -label:bug.