Can't run the converted model with vllm.LLM #32
Description
I converted the deepseek-r1-distill-qwen-7b model, and when i play with the converted mdoels using vllm.LLM as you said in the readme.md,
it can't work, and the error message is TypeError: TypeError: MLACommonImpl.__init__() got an unexpected keyword argument 'rotary_emb'.