flagos-ai/vllm-plugin-FL

Python 15 stars

A vLLM plugin built on the FlagOS unified multi-chip backend.

✓ Synced 5h ago Share on X →

README badge: [![ngmi](https://ngmi.review/badge/flagos-ai/vllm-plugin-FL.svg)](https://ngmi.review/repo/flagos-ai/vllm-plugin-FL)

36 Merged PRs

3 days Avg Merge Time

0m Fastest PR

1 month Slowest PR

      #237
      Global Speed Rank
    

Top Reviewers

#	Title	Author	Time	Reviews	Blocks
#60	Update the readme	@aoyulong	0m	1	✓
#59	Update README.md	@aoyulong	0m	1	✓
#57	Update the readme	@aoyulong	9.1h	5	✓
#56	feat: support qwen3.5	@yixiaodapeng	1 day	2	✓
#50	Add basic CI workflow	@xmhubj	2 days	2	✓
#49	update qwen3next to use fused_recurrent in flag_gems	@ceci3	20.9h	2	✓
#39	Add unit and functional tests	@xmhubj	6 days	7	✓
#45	update install method of flagcx torch plugin	@mikethegoblin	23.5h	2	✓
#48	add csrc	@ceci3	15m	2	✓
#43	Support Instance Context Passing in Dispatch System	@xin2an	1 day	9	✓
#42	update readme	@yixiaodapeng	21m	2	✓
#33	explicit cleanup engine core	@yangrudan	5 days	7	✓
#7	fix vllm_fl/utils.py and README.md error	@taozhiwei	1 month	2	✓
#38	Merge main to dev	@ceci3	52m	2	✓
#32	Support Kimi-K2.5 cooperated with claude under vllm 0.13.0	@shh2000	2 days	6	✓
#34	Add platform-specific operator config (ascend.yaml/cuda.yaml) with env var override support and enable FlagGems attention backend	@xin2an	1 day	7	✓
#35	Add pytest correctness and performance testing	@luoyc123	23.3h	5	✓
#36	[bugfix] fix import error	@cyber-pioneer	5m	2	✓
#17	[New Feature] Add op autotune over all backends	@cyber-pioneer	15 days	5	✓
#26	add minicpm	@ceci3	1 day	3	✓