-
Notifications
You must be signed in to change notification settings - Fork 114
Pull requests: ROCm/ATOM
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[atom] Add diffusion subsystem and MiniMax-H3 video+audio generation
#1836
opened Aug 8, 2026 by
carlushuang
Collaborator
Loading…
fix(scheduler): unblock decode behind a long chunked prefill
ci:atom
#1835
opened Aug 7, 2026 by
hippothewild
Loading…
1 task
[ATOM SGL] [bug fix]fix glm52 regression
#1834
opened Aug 7, 2026 by
ZhiweiYan-96
Contributor
Loading…
1 task
[gfx1250] Route wo_a grouped LoRA through the flydsl a8w4 batched GEMM
#1833
opened Aug 7, 2026 by
XingerZhu
Loading…
4 tasks
fix(attention): guard sparse MLA index conversion
ci:full
#1831
opened Aug 7, 2026 by
qichu-yun
Contributor
Loading…
1 task
[kimi-k3] attn_res: add opt-in aiter attn_res_gate routing via ATOM_U…
#1830
opened Aug 7, 2026 by
yanxuer-999
•
Draft
1 task
[Atomesh][CI] guard SWE-bench disk usage and report prefix-cache hit rate
#1826
opened Aug 7, 2026 by
Phi-C
Contributor
Loading…
1 task
[ATOM SGL][model] Kimi K3
ci:sglang
#1821
opened Aug 6, 2026 by
ZhiweiYan-96
Contributor
Loading…
1 task
[ATOM][Kimi-K3] add ATOM_USE_AITER_KDA to run KDA prefill on aiter
#1820
opened Aug 6, 2026 by
Liang-jianhao97
Contributor
Loading…
1 task done
feat(glm): enable persistent decode for GLM-5.2 DPA
#1816
opened Aug 6, 2026 by
wanzhenchn
Contributor
Loading…
fix(glm-dsa): enable stable top-k for tensor parallel
#1811
opened Aug 6, 2026 by
yhl-amd
Contributor
Loading…
[fix](dsv4): fix dsv4 accuracy error
ci:vllm
#1809
opened Aug 6, 2026 by
PerryZhang01
Contributor
Loading…
1 task
feat: add reliable DP/TP collective RPC support for ATOM workers
#1808
opened Aug 6, 2026 by
JiaoliangYu
Contributor
•
Draft
1 task
[DSpark] Support kimi k3 Dspark PS MLA and draft model casual False
#1806
opened Aug 6, 2026 by
ZhangLirong-amd
Collaborator
Loading…
1 task
[kimi-k3] DSpark speculative decoding on vLLM 0.26
#1805
opened Aug 5, 2026 by
whx-sjtu
Contributor
Loading…
fix: include content="" in streaming role chunk to match OpenAI/vLLM shape
ci:atom
#1803
opened Aug 5, 2026 by
shantipriya-amd
Loading…
1 task done
[SGL+ATOM]fix[plugin]: guard GLM flat fmoe on MI308
ci:sglang
#1798
opened Aug 5, 2026 by
zhangxinyuanliuhengyu
Contributor
Loading…
[feat](vllm): upgrade vllm to 0.26.1
#1794
opened Aug 5, 2026 by
zejunchen-zejun
Collaborator
•
Draft
[vLLM-ATOM] Support DSpark speculative decoding for Kimi-K3
ci:vllm
#1789
opened Aug 4, 2026 by
whx-sjtu
Contributor
Loading…
4 of 5 tasks
optimize attn res ref fla and fuse all possible part around it
ci:atom
#1788
opened Aug 4, 2026 by
ganyi1996ppo
Contributor
Loading…
1 task
Previous Next
ProTip!
What’s not been updated in a month: updated:<2026-07-07.