Skip to content

Pull requests: jd-opensource/xllm

New pull request New
71 Open 1,026 Closed
71 Open 1,026 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

feat: add npu_recurrent_gated_delta_rule and chunk_gated_delta_rule fusion operators for qwen3.5/qwen3-next.
#1262 opened Apr 11, 2026 by fems14 • Draft
5
bugfix: init ssm_cache by config ssm_cache_type and unify compute precision to fp32.
#1259 opened Apr 10, 2026 by JC-ut0 Contributor Loading…
7
bugfix: fix DeepSeek V3 crash and V3.2 prefix-cache OOM.
#1258 opened Apr 10, 2026 by DongheJin Collaborator Loading…
2
bugfix: fix CFG negative prompt judgment & simplify variable names fo…
#1256 opened Apr 10, 2026 by yiming-l21 Collaborator Loading…
feat: support Qwen down_proj fallback for compressed-tensors ignored modules.
#1254 opened Apr 10, 2026 by yingxudeng Collaborator Loading…
3
bugfix: remove spurious backslash breaking output redirection in launch scripts.
#1248 opened Apr 10, 2026 by kuishou68 Loading…
1
feat: add mlu mooncake pd push support.
#1246 opened Apr 10, 2026 by phantomlei3 Collaborator Loading…
21
refactor: simplify xllm server startup routing and lifecycle helpers.
#1243 opened Apr 9, 2026 by liutongxuan Collaborator Loading…
5
perf: Qwen Image Optimize.
#1242 opened Apr 9, 2026 by shan-chen-feng Collaborator Loading…
feat: support in-batch prefix cache.
#1240 opened Apr 9, 2026 by Clement-Wang26 Collaborator Loading…
8
bugfix: optimize multi-modal preprocess accuracy.
#1235 opened Apr 9, 2026 by wly-115 Collaborator Loading…
5
feat: add configurable decode ACL-graph fallback threshold.
#1233 opened Apr 8, 2026 by DongheJin Collaborator Loading…
6
feat: support tensor parallel for Flux model on npu device.
#1231 opened Apr 8, 2026 by z-jun03 Collaborator Loading…
4
perf: Qwen image optimize.
#1230 opened Apr 8, 2026 by shan-chen-feng Collaborator Loading…
13
feat: expose startup runtime flags through c and python apis.
#1229 opened Apr 8, 2026 by RobbieLeung Collaborator • Draft
1
feat: enable rec fast sampler for llm beam search.
#1224 opened Apr 8, 2026 by RobbieLeung Collaborator Loading…
9
feat: improve cuda shared memory tensor handling.
#1222 opened Apr 8, 2026 by RobbieLeung Collaborator Loading…
5
github: update GitHub Actions workflow to handle issue comments and permissions for CI triggers.
#1221 opened Apr 8, 2026 by XuZhang99 Collaborator • Draft
2
bugfix: support per-fork disagg pd port for fork master.
#1220 opened Apr 8, 2026 by Clement-Wang26 Collaborator Loading…
3
bugfix: fix failures when EP/DP and ACL Graph are enabled simultaneously.
#1218 opened Apr 8, 2026 by DongheJin Collaborator Loading…
4
feat: support Qwen3.5-VL model on npu device[6/N].
#1212 opened Apr 7, 2026 by yingxudeng Collaborator Loading…
7
feat: add more model manual loader.
#1210 opened Apr 7, 2026 by Clement-Wang26 Collaborator Loading…
9
feat: add beam top for beam search last step for rec.
#1209 opened Apr 7, 2026 by ChrisGao001 Collaborator Loading…
12
bugfix: fix the problem of incorrect kv cache data format when enabli…
#1203 opened Apr 7, 2026 by longhui-z Contributor Loading…
3
bugfix: update Glm5-W8A8 & Draft Model on npu device.
#1193 opened Apr 6, 2026 by sanlio36 Collaborator Loading…
5
ProTip! Filter pull requests by the default branch with base:main.