Skip to content

Pull requests: ROCm/ATOM

New pull request New
92 Open 588 Closed
92 Open 588 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

Add DeepSeek-V3.2 fused indexer path
#788 opened May 14, 2026 by XiaobingSuper Contributor Loading…
7
[fix](gpt-oss): fix quark quantized model in moe bias
#787 opened May 14, 2026 by PerryZhang01 Contributor Loading…
Add DSR1-MXFP4 recipe for MI355X (Team Jons contest submission, 2840/3000)
#786 opened May 14, 2026 by j0ons Loading…
2
ci: branch-aware docker release + fix benchmark model selection
#785 opened May 14, 2026 by ZhangLirong-amd Contributor Loading…
1 task
3
fix prefix default on
#784 opened May 14, 2026 by jiayyu Contributor Loading…
1 task
1
fix: qwen_3.5 bf16 accurancy
#782 opened May 14, 2026 by JiaoliangYu Loading…
1 task
ci(benchmark): upgrade Kimi K2.5 to K2.6
#781 opened May 14, 2026 by carlushuang Contributor Loading…
1 of 2 tasks
[vLLM Plugin] Upgrade vLLM version to v0.21.0
#780 opened May 14, 2026 by whx-sjtu Contributor • Draft
2 of 6 tasks
[codex] DeepSeek FP4 MTP decode safeguards and MLA hooks
#779 opened May 13, 2026 by josusanmartin • Draft
feat(server): add Anthropic Messages API endpoint (/v1/messages)
#778 opened May 13, 2026 by carlushuang Contributor Loading…
4 of 5 tasks
(ci)[SGLang-ATOM]: Add Qwen3.5 cases for ci, nightly and benchmark
#777 opened May 13, 2026 by zhuyuhua-v Collaborator Loading…
8
Qwen3Next MTP for vLLM plugin mode
#772 opened May 13, 2026 by ganyi1996ppo Contributor Loading…
1 task
14
Add mooncake dockerfile build
#771 opened May 13, 2026 by ZhangLirong-amd Contributor Loading…
1 task
22
[Perf][vLLM-ATOM] Optimize Sparse MLA in vLLM-ATOM
#765 opened May 12, 2026 by kliuae Contributor Loading…
1 task
1
[MoE] adapt to triton_kernels matmul_ogs -> matmul rename
#763 opened May 12, 2026 by Liang-jianhao97 Loading…
1 task done
perf: flydsl gdr decode default on
#752 opened May 11, 2026 by xytpai Contributor Loading…
[feat][Attention Refactor] Reconstruct the Attention arch
#750 opened May 11, 2026 by zejunchen-zejun Collaborator • Draft
18
Add Mistral-3-8B + Qwen3-8B-FP8 + native triton attention backend for gfx1201 (RDNA4 / RX 9070 XT)
#749 opened May 11, 2026 by carlushuang Contributor Loading…
1
[feat][breaking] Enable prefix caching by default
#741 opened May 11, 2026 by functionstackx Contributor Loading…
3 of 6 tasks
1
8
Fpz/chunk prefill
#740 opened May 11, 2026 by jiayyu Contributor Loading…
1 task
3
perf: optimize GDN decode with SGLang fused recurrent kernel
#727 opened May 9, 2026 by zovonoir Contributor Loading…
1 of 2 tasks
6
[Feat] Support GLM-4.7 MTP in vLLM-ATOM plugin
#722 opened May 8, 2026 by kliuae Contributor Loading…
1 task
5
docs: deploy compressor page with docs workflow
#715 opened May 7, 2026 by gyohuangxin Member Loading…
2
perf: fused Triton kernels for Qwen3.5 RMSNorm and MRoPE
#708 opened May 7, 2026 by zovonoir Contributor Loading…
1 of 2 tasks
14
[ci] add Qwen3.5 Dense/MoE models accuracy validation and benchmark tests for atom-plugined sglang
#700 opened May 6, 2026 by wanzhenchn Contributor Loading…
3
ProTip! Follow long discussions with comments:>50.