fangzheng

I have same problem, and I find the reason is that the opencv verion provided in zzopencv.sh is too low(4.3.0), so downloading opencv 4.5.2 and recompiling opencv 4.5.2 and denseflow...

> 看一下 #46 你好，我看了那个问题，但是我这里是用的vllm加速推理。代码基本如下 from vllm import LLM,SamplingParams model=LLM('qwen_0.5b_finetuned') tokenizer=AutoTokennizer.from_pretrained('qwen_0.5b_finetuned') sampling_params=SamplingParams(top_p=0.8,temperature=0.5,max_token=5) prompt='你好' messages=[ {"role":"system","content":"You are a helpful assistant"} {"role":"user","content":prompt} ] text=tokenizer.apply_chat_template(messages,tokenize=False,add_generation_prompt=True) outputs=model.generate([text],sampling_params) 上述的代码输出为：lon euraszeich deleting。但是我用没微调过的输出就是正常的，你看我这个应该在哪里改参数可以使它输出正确呢？