vllm-project/vllm

main

vllm

Grade	Name	Complexity	Churn	Issues
C	vllm\model_executor\models\qwen2_5_vl.py	89	1	7
C	vllm\model_executor\models\qwen3_omni_moe_thinker.py	171	1	11
C	vllm\models\deepseek_v4\nvidia\model.py	120	2	8
C	vllm\models\deepseek_v4\xpu\model.py	83	1	17
D	csrc\libtorch_stable\attention\mla\cutlass_sm100_mla\kernel\sm100_fmha_mla_tma_warpspecialized.hpp	182	1	21
D	tests\models\multimodal\generation\vlm_utils\model_utils.py	205	1	22
D	vllm\model_executor\models\qwen3_vl.py	181	1	14
F	.buildkite\scripts\publish-release-images.sh		1	88
F	csrc\libtorch_stable\moe\marlin_moe_wna16\marlin_template.h	1	1	17
F	csrc\libtorch_stable\quantization\marlin\marlin_template.h	1	1	16
F	examples\generate\multimodal\vision_language_multi_image_offline.py	18	1	26
F	examples\generate\multimodal\vision_language_offline.py	110	1	16
F	tests\kernels\moe\test_moe_layer.py	149	1	41
F	tests\lora\conftest.py	4	1	28
F	vllm\third_party\pynvml.py	172	1	883
F	vllm\v1\sample\ops\topk_topp_triton.py	91	1	270
F	vllm\v1\worker\gpu_model_runner.py	806	6	15