Skip to content

Pull requests: jd-opensource/xllm

New pull request New
83 Open 1,282 Closed
83 Open 1,282 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

perf(npu): eliminate redundant Transpose in Qwen3.5 MTP spec verify conv path
#1536 opened May 23, 2026 by pjgao Loading…
7
feat: update mlu container to 26.04.
#1535 opened May 23, 2026 by phantomlei3 Collaborator • Draft
4
bugfix: drive pd push transfer by cursor to avoid miss transfer.
#1532 opened May 22, 2026 by phantomlei3 Collaborator • Draft
1
feat: support interlayer add norm and SplitRmsnormRope operation for qwen3.
#1531 opened May 22, 2026 by shan-chen-feng Collaborator Loading…
5
refactor: refact multimodal processor.
#1530 opened May 22, 2026 by wly-115 Collaborator Loading…
4
bugfix: resolve causal_conv1d tiling failure for qwen3.5 gdn decode with padded batches (#1428 cherry-pick to main)
#1529 opened May 22, 2026 by pjgao Loading…
3
feat: support MiMo-7B-Base on cuda device.
#1523 opened May 22, 2026 by Dragonliu2018 Contributor Loading…
20
feat: add dcu backend support.
#1522 opened May 22, 2026 by WenQ7 Loading…
23
feat: support dumping xllm server flags to json file.
#1518 opened May 22, 2026 by XuZhang99 Collaborator Loading…
2
feat:remove unused func and support deepseek_v4_mtp graph on npu.
#1517 opened May 22, 2026 by panxua Contributor Loading…
7
feat: expose cached token usage in responses.
#1514 opened May 21, 2026 by zhang-minchao Collaborator Loading…
3
refactor: remove xattention one-stage decode path.
#1504 opened May 21, 2026 by LMX-xin Collaborator • Draft
2
feat: enable REC XAttention for Qwen3 MoE on cuda device.
#1500 opened May 20, 2026 by LMX-xin Collaborator Loading…
1
feat: support vae parallel for qwen-image-edit-plus.
#1499 opened May 20, 2026 by shan-chen-feng Collaborator Loading…
13
feat: add TileLang chunk_gated_delta_rule_fwd_h kernel.
#1498 opened May 20, 2026 by fengz72 Loading…
5
bugfix: use max_concurrent_requests for single block and linear state allocation.
#1496 opened May 20, 2026 by pjgao Loading…
7
feat: support customized multimodal preprocess configs.
#1481 opened May 19, 2026 by xanecdotex Collaborator Loading…
4
refactor: remove negative condition when choosing decode or prefill
#1475 opened May 18, 2026 by rauletorresc Contributor Loading…
feat: parallelize multimodal decode in request transfer.
#1474 opened May 18, 2026 by wly-115 Collaborator Loading…
6
refactor: split forward inputs from model input params [3 / 3].
#1469 opened May 18, 2026 by RobbieLeung Collaborator Loading…
15
bugfix: ensure tensor contiguous layout before protobuf serialization.
#1468 opened May 18, 2026 by a120092009 Collaborator Loading…
4
fused_sigmoid_gating_tilelang tilelang adapt in qwen3.x
#1465 opened May 15, 2026 by BikingNow Loading…
10
feat: Support Flux2 in other common components
#1463 opened May 15, 2026 by wang-shuibin Loading…
7
bugfix: reduce acl graph memory overhead.
#1457 opened May 15, 2026 by RobbieLeung Collaborator Loading…
2
docs: exporting a draft model from a quantized model.
#1455 opened May 14, 2026 by rauletorresc Contributor Loading…
8
ProTip! Follow long discussions with comments:>50.