-
Notifications
You must be signed in to change notification settings - Fork 195
Pull requests: jd-opensource/xllm
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
feat: support remote mtp bootstrap for pd mlu decode.
#1396
opened May 6, 2026 by
phantomlei3
Collaborator
Loading…
feat: add automatic numa binding for mlu.
#1395
opened May 6, 2026 by
phantomlei3
Collaborator
Loading…
feat: support wan22's text_encoder and scheduler.
#1393
opened May 6, 2026 by
ethan686
Contributor
Loading…
feat: support chunked prefill for mlu pd.
#1392
opened May 6, 2026 by
phantomlei3
Collaborator
Loading…
bugfix: fix REC xattention without tensor core support.
#1391
opened May 2, 2026 by
LMX-xin
Collaborator
Loading…
feat: support deepseek-v4 swiglu_limit on npu.
#1390
opened Apr 30, 2026 by
yingxudeng
Collaborator
Loading…
refactor: move all ut to tests dir and add ut skill.
#1389
opened Apr 29, 2026 by
XuZhang99
Collaborator
Loading…
bugfix: fix chunkedprefill error for cuda device.
#1387
opened Apr 29, 2026 by
weizhehuang0827
Collaborator
Loading…
bugfix: add npu_kernel_backend param for offline inference to fix tp bug.
#1386
opened Apr 29, 2026 by
weizhehuang0827
Collaborator
Loading…
bugfix: fix MoE buffer and shared expert quant.
#1381
opened Apr 29, 2026 by
DongheJin
Collaborator
Loading…
feat: pass vlm tools and chat template kwargs to prompt rendering.
#1380
opened Apr 29, 2026 by
yingxudeng
Collaborator
Loading…
feat: add qwen3.5 for mlu backend without GDN.
#1377
opened Apr 29, 2026 by
phantomlei3
Collaborator
Loading…
refactor: add model input param groups and factory.
#1366
opened Apr 28, 2026 by
liutongxuan
Collaborator
Loading…
feat: add triton npu runtime asset staging.
#1361
opened Apr 27, 2026 by
yingxudeng
Collaborator
Loading…
refactor: optimize speed of speculative decoding.
#1356
opened Apr 27, 2026 by
RobbieLeung
Collaborator
Loading…
feat: update qwen3_gated_delta_net_base.cpp for NPU torch layer
#1340
opened Apr 24, 2026 by
Sinle4Cat
Loading…
feat: add tilelang fast path for acl graph decode metadata update.
#1337
opened Apr 23, 2026 by
zhang-minchao
Collaborator
•
Draft
refactor: add model input param groups and factory.
#1331
opened Apr 22, 2026 by
liutongxuan
Collaborator
Loading…
bugfix: fix precision problem for QwenImageEdit.
#1327
opened Apr 22, 2026 by
xiao-yu-chen
Collaborator
Loading…
feat: support multi-priority scheduling for PD disagg.
#1318
opened Apr 21, 2026 by
weizhehuang0827
Collaborator
Loading…
feat: add rmsnorm_qk_rope and rmsnorm_qk for ilu device.
#1317
opened Apr 21, 2026 by
laneeeee
Contributor
Loading…
Previous Next
ProTip!
Type g i on any issue or pull request to go back to the issue listing page.