Repository Issues
vllm-project/vllm
A high-throughput and memory-efficient inference and serving engine for LLMs
Issues
Open
[Feature]: Remove attention layer name from `unified_kv_cache_update`
feature requestgood first issuehelp wantedstalestartup-uxtorch.compile
Has a beginner-friendly labelContributing guide available
8 comments1 reaction2 assignees
Closed
[Feature][Help Wanted]: Add tuning script and config files for Mamba selective_state_update kernel
feature requestgood first issuehelp wanted
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
11 comments0 reactions0 assignees
Closed
[Bug][Help Wanted]: PPLX + vLLM CUTLASS FP8 Gives Incorrect Responses
bughelp wantedstale
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
6 comments0 reactions0 assignees
Closed
[Feature]: Integrate RMS+fp4 fused kernel from FlashInfer
feature requestgood first issuehelp wantedstaletorch.compile
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
14 comments0 reactions1 assignee
Open
[Bug]: Wrong timestamps if audio > 30s
buggood first issuehelp wanted
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
11 comments1 reaction2 assignees
Open
[Feature]: Extract KV-Cache update from all attention backends
feature requestgood first issuehelp wanted
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
51 comments0 reactions9 assignees
Open
[Feature]: Refactor Int8ScaledMMLinearLayerConfig to use QuantKey
feature requestgood first issuehelp wanted
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
7 comments0 reactions1 assignee
Closed
[Feature]: Support DCP with FP8 KV Cache
feature requesthelp wanted
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
4 comments0 reactions0 assignees
Closed
[Bug]: 0% Acceptance rate with FI Cutlass DeepSeekR1 NVFP4 with mtp ep
bughelp wantedstale
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
4 comments0 reactions0 assignees
Open
[Feature]: Use kernel abstraction for fp4 scaled mm
feature requesthelp wantedunstale
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
12 comments0 reactions1 assignee
Open
[Feature]: Error Logging Redesign
feature requesthelp wanted
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
7 comments0 reactions1 assignee
Open
[Bug]: ModelOpt Llama-4 Checkpoints Take 5+ minutes to load
bugfeature requestgood first issuehelp wanted
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
12 comments0 reactions0 assignees
Open
[Feature]: Enable LoRA support for tower and connector in more MM models
feature requesthelp wantedmulti-modality
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
10 comments2 reactions0 assignees
Open
[Feature][Cleanup]: Unify `vllm.utils.flashinfer` and `vllm.model_executor.layers.quantization.utils.flashinfer_utils`
feature requestgood first issuehelp wanted
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
7 comments0 reactions0 assignees
Closed
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
3 comments0 reactions0 assignees
Open
[RFC]: Improve environment variable declaration and handling
RFCgood first issuehelp wantedkeep-open
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
9 comments2 reactions1 assignee
Closed
[Bug]: FlashInfer Incompatible with Sleep Mode
bughelp wantedstale
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
4 comments2 reactions0 assignees
Open
[Feature]: Default eplb num_redundant_experts to the lowest valid value if unspecified
feature requestgood first issuehelp wantedunstale
Why recommendedNo assignee yet · Has a beginner-friendly label
No assignee yetHas a beginner-friendly labelContributing guide available
9 comments0 reactions0 assignees
Open
[Feature]: Add support for fused fp8 output to FlashAttention 3
feature requesthelp wantedkeep-openperformancetorch.compile
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
15 comments0 reactions1 assignee
Open
[Feature]: Qwen3 Omni Transcriptions
feature requestgood first issue
Why recommendedHas a beginner-friendly label · Contributing guide available
Has a beginner-friendly labelContributing guide available
9 comments0 reactions1 assignee