vllm-project/vllm

main

vllm

Grade	Name	Complexity	Churn	Issues
A	vllm\tokenizers\deepseek_v4_encoding.py	84	1	1
A	vllm\model_executor\layers\fused_moe\oracle\fp8.py	83	1	2
A	vllm\distributed\kv_transfer\kv_connector\v1\nixl\push_worker.py	83	2	2
A	vllm\model_executor\layers\quantization\utils\marlin_utils_fp4.py	83	1	1
A	tests\tool_parsers\test_step3p5_tool_parser.py	83	1	-
C	vllm\models\deepseek_v4\xpu\model.py	83	1	17
A	benchmarks\attention_benchmarks\mla_runner.py	83	1	2
B	vllm\model_executor\models\qwen2_vl.py	82	1	6
A	tests\v1\core\test_prefix_caching.py	82	1	-
A	vllm\entrypoints\openai\responses\context.py	82	1	-
A	vllm\model_executor\layers\quantization\moe_wna16.py	81	1	3
A	csrc\cpu\spec_decode_utils.cpp	81	1	-
A	vllm\model_executor\models\registry.py	81	5	4
A	vllm\model_executor\models\transformers\base.py	80	1	2
A	csrc\cpu\sgl-kernels\moe.cpp	80	1	5
A	vllm\platforms\cpu.py	80	1	3
A	benchmarks\kernels\benchmark_lora.py	80	1	-
A	vllm\distributed\device_communicators\shm_broadcast.py	80	2	2
A	tests\v1\e2e\general\test_mamba_prefix_cache.py	79	1	2
A	vllm\distributed\kv_transfer\kv_connector\v1\moriio\moriio_engine.py	79	1	-