-
Notifications
You must be signed in to change notification settings - Fork 63
Pull requests: ROCm/ATOM
Author
Label
Projects
Milestones
Reviews
Assignee
Sort
Pull requests list
[ATOM-vllm-benchmark] Debug 'no such file or directory benchmark_matrix.json'
#994
opened May 31, 2026 by
junyyang-amd
Contributor
Loading…
ci(accuracy): set Qwen3.5-35B-A3B TP2 baseline to 0.85
#993
opened May 31, 2026 by
JiaoliangYu
Contributor
Loading…
1 task
[atom-sgl-benchmark] Debug benchmark timeout
#977
opened May 29, 2026 by
junyyang-amd
Contributor
Loading…
[sgl][recipe]update sglang-atom ds fp4 recipe
#975
opened May 29, 2026 by
zhuyuhua-v
Collaborator
•
Draft
[Feat] Fused qknorm + quant for dpsk v2 model
#963
opened May 28, 2026 by
qichu-yun
Contributor
Loading…
1 task
[Feat] enable fuse share expert in DeepSeek-R1-0528-MXFP4-v2 in sgl atom
#958
opened May 28, 2026 by
ZLkanyo009
Contributor
Loading…
[Bugfix] eagle: fix MTP draft warmup crash on mrope-config target models
#957
opened May 28, 2026 by
yhl-amd
Contributor
Loading…
3 tasks
Adapt EAGLE 3.1 (fc_norm + post-norm feedback) and add Kimi K2.6 MLA draft
#943
opened May 27, 2026 by
yhl-amd
Contributor
Loading…
3 of 5 tasks
[Fix] Enable dpsk r1 mxfp4 V2 model
#934
opened May 26, 2026 by
qichu-yun
Contributor
Loading…
1 task
feat: add RTP-LLM plugin integration and Qwen3.5 MoE support
#925
opened May 26, 2026 by
zhaoan12-prc
•
Draft
[Atomesh] PD disaggregation router with multi-node support
#919
opened May 26, 2026 by
Jasen2201
Loading…
fix(plugin): size MTP decode scratch by token capacity
#916
opened May 25, 2026 by
san-tian
Loading…
[ATOM-SGLang] Add benchmark and ci for mi308.
#909
opened May 25, 2026 by
junyyang-amd
Contributor
Loading…
[fix][attn] fail fast when --page-size < kv element width
#905
opened May 25, 2026 by
zovonoir
Contributor
Loading…
Previous Next
ProTip!
Type g p on any issue or pull request to go back to the pull request listing page.