OK   torch: 2.4.0
OK   transformers: 4.49.0
OK   Qwen2.5VL_model: OK
OK   AutoProcessor: OK
OK   accelerate: 1.13.0
OK   deepspeed: 0.18.7
OK   peft: 0.18.1
OK   flash_attn: 2.8.3
FAIL ms_swift: No module named 'swift.llm'
OK   PIL: OK
OK   qwen_vl_utils: OK
OK   CUDA: 8 GPUs