vllm-project/vllm

main

vllm

Grade	Name	Complexity	Churn	Issues
A	vllm\v1\core\kv_cache_utils.py	189	3	-
B	vllm\v1\spec_decode\llm_base_proposer.py	183	3	8
D	csrc\libtorch_stable\attention\mla\cutlass_sm100_mla\kernel\sm100_fmha_mla_tma_warpspecialized.hpp	182	1	21
D	vllm\model_executor\models\qwen3_vl.py	181	1	14
A	vllm\multimodal\video.py	180	1	5
A	vllm\config\speculative.py	174	3	3
F	vllm\third_party\pynvml.py	172	1	883
C	vllm\model_executor\models\qwen3_omni_moe_thinker.py	171	1	11
A	vllm\distributed\kv_transfer\kv_connector\v1\mooncake\store\worker.py	170	4	1
C	vllm\model_executor\models\glm4_1v.py	167	2	15
A	vllm\v1\engine\core_client.py	166	1	1
A	vllm\model_executor\model_loader\weight_utils.py	166	1	2
A	vllm\model_executor\models\phi4mm_utils.py	165	1	2
A	vllm\model_executor\models\molmo2.py	164	1	5
A	vllm\distributed\kv_transfer\kv_connector\v1\offloading\scheduler.py	162	1	3
A	.buildkite\performance-benchmarks\scripts\compare-json-results.py	161	1	3
A	vllm\model_executor\layers\attention\mla_attention.py	160	1	3
A	vllm\tool_parsers\utils.py	160	1	3
A	benchmarks\attention_benchmarks\benchmark.py	160	1	2
A	vllm\benchmarks\throughput.py	159	1	5