Skip to content

Pull requests: jd-opensource/xllm

New pull request New
37 Open 719 Closed
37 Open 719 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: support index cache transfer in PD separate deployment scenario.
#895 opened Feb 7, 2026 by sunbaosong Loading…
3
feat: adapt for CANN 8.5 and PyTorch 2.7.1 for npu device.
#891 opened Feb 6, 2026 by haimbb000 Loading…
5
feat: add beam search for recommandation in cuda env.
#890 opened Feb 6, 2026 by liujinguang0125 Loading…
11
bugfix: fix mrope position tensor shape mismatch in NPU graph mode.
#885 opened Feb 4, 2026 by QwertyJack Loading…
2 tasks done
1
15
feat: add DeepSeek-V3.2 MTP on NPU.
#879 opened Feb 4, 2026 by edison240121 Loading…
3
feat: support mm_data rpc transport.
#876 opened Feb 3, 2026 by wly-115 Loading…
3
feat: add VMM submitter APIs for non-blocking vmm::map/unmap.
#874 opened Feb 3, 2026 by shifengmin Loading…
47
feat: add piecewise cudagraph for prefill.
#872 opened Feb 3, 2026 by zhang-minchao Loading…
28
docs: add iluvatar model support list.
#871 opened Feb 3, 2026 by laneeeee Loading…
1
feat: add optimization codes for qwen3 moe
#865 opened Feb 3, 2026 by panxua Loading…
2
bugfix: fix early return in prefetch_from_storage.
#862 opened Feb 2, 2026 by shifengmin Loading…
2
feat: dynamic and scalable multi-model support.
#861 opened Feb 2, 2026 by Clement-Wang26 Loading…
5
feat: support eagle3 for qwen3 series.
#859 opened Feb 2, 2026 by RobbieLeung Loading…
8
feat: add kernels that support Qwen3 model on musa device.
#856 opened Feb 2, 2026 by FleckyFelix Loading…
6
feat: support LongCat-Image on cuda device.
#849 opened Jan 31, 2026 by Dragonliu2018 Loading…
20
refactor: recfactor xllm docs arch for better readability.
#845 opened Jan 31, 2026 by XuZhang99 • Draft
feat: implement rpc interface in APIService for xllm service internal usage.
#837 opened Jan 30, 2026 by weizhehuang0827 Loading…
bugfix: fix MTP k>1 crash by loading embed_tokens weights
#836 opened Jan 29, 2026 by QwertyJack • Draft
3 tasks done
1
6
feat: support QwenImageEditPlus pipeline with embedding infer.
#834 opened Jan 29, 2026 by shan-chen-feng Loading…
10
bugfix: fix KV cache memory leak when prefix cache is enabled.
#829 opened Jan 29, 2026 by QwertyJack • Draft
3 tasks
1
9
feat: support mtp graph for mlu graph executor.
#825 opened Jan 28, 2026 by a120092009 Loading…
6
bugfix: get suitable token budget that allocated for sequence when enable MTP with overlap.
#823 opened Jan 28, 2026 by RobbieLeung Loading…
4
bugfix: fix missing M-RoPE section in GLM-4V model args.
#822 opened Jan 28, 2026 by wly-115 Loading…
6
feat: implement bidirectional remote host to local device kv cache transfer and batch offload.
#812 opened Jan 27, 2026 by Kang-Meng • Draft
5
refactor: centralize ffmpeg resource cleanup.
#810 opened Jan 27, 2026 by xanecdotex Loading…
6
ProTip! Add no:assignee to see everything that’s not assigned.