Skip to content

Pull requests: ROCm/ATOM

New pull request New
70 Open 456 Closed
70 Open 456 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 Qwen3.5 FP4 to vLLM-ATOM nightly accuracy check and benchmark
#593 opened Apr 17, 2026 by wuhuikx Collaborator • Draft
1 task
3
docker: pin base image from ubuntu 24.04 py3.12 to ubuntu 22.04 py3.10
#592 opened Apr 17, 2026 by Jasen2201 Loading…
4
Update mori version and make tbo in nightly
#590 opened Apr 17, 2026 by ZhangLirong-amd Contributor Loading…
1 task
4
[qwen3next] fix gdr
#589 opened Apr 17, 2026 by ganyi1996ppo Contributor Loading…
1 task
2
fix kv cache fp8 issue
#588 opened Apr 17, 2026 by ganyi1996ppo Contributor Loading…
1 task
1
[dashboard] show docker info in dashboard accuracy chart
#587 opened Apr 17, 2026 by zejunchen-zejun Collaborator Loading…
1
ci: add Qwen3.5-397B-A17B-MXFP4 to per-PR accuracy test
#586 opened Apr 16, 2026 by valarLip Collaborator Loading…
1 task
1
[Acc] remove flatten in dpsk fp4 by quantizing mla weights && a…
#582 opened Apr 16, 2026 by ZLkanyo009 Contributor Loading…
2
[atom-vllm] by default disable vllm mori because the atom mori will be used for atom-vllm
#579 opened Apr 16, 2026 by zejunchen-zejun Collaborator • Draft
2
[Gluon] [Triton] [MI450] [MI350] Enable Triton/Gluon MLA with block_size 64 preshuffled kv_buffer option for decode
#578 opened Apr 15, 2026 by k50112113 Contributor Loading…
[fix](metadata): optimize attention metadata
#571 opened Apr 15, 2026 by PerryZhang01 Contributor Loading…
2
[Gluon] [Triton] [MI450] [MI350] Enable Unified Attention option for decode
#566 opened Apr 14, 2026 by k50112113 Contributor Loading…
MiMo-V2-Flash Support
#560 opened Apr 14, 2026 by wufann Loading…
1 task
3
[Feat][Plugin] Enable spec decoding for GLM5 (vLLM Plugin)
#557 opened Apr 14, 2026 by whx-sjtu Contributor • Draft
1 task done
3
CI: make ATOM test workflow reusable
#554 opened Apr 14, 2026 by gyohuangxin Member Loading…
1 of 5 tasks
6
[feat] Add RLHF rollout integration support (verl)
#549 opened Apr 13, 2026 by sijyang Loading…
5
feat: add Gemma4 31B support for standalone and vLLM plugin mode on MI355X
#546 opened Apr 12, 2026 by ClementLinCF Loading…
14
[feat](minimax): support minimax-2.5 in atom-vllm mode
#545 opened Apr 12, 2026 by PerryZhang01 Contributor Loading…
5
Update the naming of vLLM-ATOM path
#541 opened Apr 10, 2026 by wuhuikx Collaborator • Draft
19
[Draft] Add vllm-omni plugin for Diffusion models Qwen Image and Wan22
#539 opened Apr 10, 2026 by tjtanaavllm • Draft
1 task
[atom-vllm][DP/EP] enable DP/EP for atom-vllm
#533 opened Apr 9, 2026 by zejunchen-zejun Collaborator • Draft
5 of 7 tasks
[feat] Add support for Qwen3.5 and Qwen3-Next to ATOM-plugined SGLang
#532 opened Apr 9, 2026 by wanzhenchn • Draft
6
[BugFix][model_ops] chunk shuffle_weights to reduce load-time memory usage
#530 opened Apr 9, 2026 by whx-sjtu Contributor Loading…
2 of 3 tasks
[Plugin] [Feature] Supoort MLA q/k norm-quant fusion with SGLang + ATOM plugin for Deepseek
#528 opened Apr 9, 2026 by qichu-yun Loading…
1 task
1
[Feat] remove flatten in atom sglang mla like atom vllm mla
#525 opened Apr 9, 2026 by ZLkanyo009 Contributor Loading…
3
ProTip! Type g i on any issue or pull request to go back to the issue listing page.