Skip to content

Pull requests: jd-opensource/xllm

New pull request New
31 Open 607 Closed
31 Open 607 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

bugfix: fix mrope calculation in the multimodal situation
#746 opened Jan 16, 2026 by shan-chen-feng Loading…
2
feat: support Qwen3-Omni-Thinker on NPU device.
#744 opened Jan 16, 2026 by shan-chen-feng Loading…
5
bugfix: fix acl_graph_executor not handling q_cu_seq_lens parameter for deepseekv3.2.
#742 opened Jan 16, 2026 by zhang-minchao Loading…
4
perf: remove useless indexer memory ops.
#741 opened Jan 16, 2026 by a120092009 Loading…
4
feat: complex message support
#740 opened Jan 15, 2026 by QwertyJack Loading…
47
bugfix: enforce tool_choice parameter to control tool calling behavior
#737 opened Jan 15, 2026 by QwertyJack Loading…
3
feat: add compile config for musa device
#734 opened Jan 15, 2026 by FleckyFelix Loading…
7
cicd: add cicd for iluvatar device
#731 opened Jan 15, 2026 by kalashnlkov Loading…
8
bugfix: fix incorrect async implementation in rerank interface.
#728 opened Jan 15, 2026 by RobbieLeung Loading…
6
bugfix: return HTTP 400 instead of crashing when Content-Length header is missing
#727 opened Jan 14, 2026 by QwertyJack Loading…
45
feat: add rec kernel and builder for pure device pipeline[2/3].
#723 opened Jan 14, 2026 by LMX-xin Loading…
3 tasks
6
bugfix: fix streaming tool call missing function name for GLM-4.7
#722 opened Jan 14, 2026 by QwertyJack Loading…
5 tasks done
feat: support deepseek mla fused_mla_q/fused_mla_kv on mlu device.
#714 opened Jan 14, 2026 by a120092009 Loading…
9
feat: add Cutlass support for Qwen3 W8A8 quantization.
#701 opened Jan 13, 2026 by yingxudeng • Draft
16
feat: support audio modal input & refactor media decoder.
#682 opened Jan 8, 2026 by xanecdotex Loading…
16
feat: implement bidirectional remote host to local device kv cache transfer and batch offload.
#663 opened Jan 7, 2026 by Kang-Meng • Draft
18
feat: auto rebase PR branch onto target branch before build.
#654 opened Jan 6, 2026 by yingxudeng • Draft
1
feat: support prefix cache for multi-modal model.
#637 opened Jan 4, 2026 by wly-115 • Draft
10
feat: support Qwen3 128k context via YaRN scaling.
#624 opened Dec 30, 2025 by yingxudeng Loading…
5
cicd: add cicd for cuda device.
#623 opened Dec 30, 2025 by XuZhang99 Loading…
4
feat: implement HCCL distributed communication for the DiT model.
#622 opened Dec 30, 2025 by z-jun03 Loading…
13
docs: add code structure documentation.
#607 opened Dec 26, 2025 by Yank000 Loading…
7
feat: support FLUX.1-Redux-dev model on npu device.
#592 opened Dec 24, 2025 by z-jun03 Loading…
24
feat: introduce USE_NPU_TORCH flag for debugging and enhance NPU support for Qwen3-Dense[4/N]. torch_npu
#590 opened Dec 23, 2025 by yingxudeng • Draft
11
feat: add xAttention support for Qwen3 generative recommendation.
#586 opened Dec 23, 2025 by LMX-xin Loading…
19
ProTip! Follow long discussions with comments:>50.