Qwen模型使用openai_api.py部署报错,RuntimeError: probability tensor contains either inf, nan or element < 0 使用Qwen仓库的openai_api.py脚本部署时会出现错误RuntimeError: probability tensor contains either inf, nan or element < 0这个问题之前也是百思不得其解,经过查阅仓库与相关仓库的issues之后的到了解决方案解决方法predict 函数内490行左右# 原代码 delay_token_num = max([len(x) for x in stop_words]) if stop_words_ids else 0 # 修改为 if stop_words: delay_token_num = max([len(x) for x in stop_words]) else: delay_token_num = 0create_chat_completion 函数,也就是/v1/chat/completions路由上,398行左右#原代码 gen_kwargs['temperature'] = request.temperature # 改为 gen_kwargs['temperature'] = request.temperature if gen_kwargs['temperature'] < 0.51: gen_kwargs['temperature'] = 0.51相关链接https://github.com/QwenLM/Qwen1.5/issues/145https://github.com/QwenLM/Qwen/issues/1121