运行example下的llm的run.py时使用chatgpt报这个错误openai.error.InvalidRequestError: 64 is greater than the maximum of 2 - 'temperature'怎么办

[kanqgg]

Describe the question

A clear and concise description of what the question is.

Environment (please complete the following information):

• OS: [e.g. mac / window]
• Python Version [e.g. 3.6]

Screenshots

If applicable, add screenshots to help explain your problem.

Additional context

Add any other context about the problem here.

[TsuMYF]

我也遇到这个问题，我发现这个问题可能发生在源代码参数传递过程中。在 ie_prompter.get_openai_result(）中可以配置temperature，max_tokens，这些参数。其中默认的temperature=0，max_tokens=64。但当我采用默认时会报错“64 is greater than the maximum of 2 - 'temperature'”.然后我修改 ie_prompter.get_openai_result(temperature=0.2），它仍报错“64 is greater than the maximum of 2 - 'temperature”。当我修改 ie_prompter.get_openai_result(temperature=0.2，max_tokens=4），它居然报错“4 is greater than the maximum of 2 - 'temperature”？？？？？是不是在源代码的参数传递过程中，错误的把max_tokens的参数当做temperature的参数传递出去了？

[GoooDte]

您好 可能是easyinstruct包里的bug 我们尽快联系相关同学修复

[TsuMYF]

@GoooDte 非常感谢，我提供我的报错记录，希望对您的修改有帮助，我也会查看自己的设置是否有些错误： Traceback (most recent call last): File "D:/pythonTry/Try2/InstructionKGC/GPT_KGC/run.py", line 45, in main result = ie_prompter.get_openai_result(temperature=0.2,max_tokens=3, File "C:\Users\hinelon\AppData\Roaming\Python\Python38\site-packages\easyinstruct\prompts\ie_prompt.py", line 276, in get_openai_result openai_result = super().get_openai_result(engine, temperature, max_tokens, top_p, frequency_penalty, presence_penalty) File "C:\Users\hinelon\AppData\Roaming\Python\Python38\site-packages\easyinstruct\prompts\base_prompt.py", line 31, in get_openai_result response = openai.Completion.create( File "C:\Users\hinelon\AppData\Roaming\Python\Python38\site-packages\openai\api_resources\completion.py", line 25, in create return super().create(*args, **kwargs) File "C:\Users\hinelon\AppData\Roaming\Python\Python38\site-packages\openai\api_resources\abstract\engine_apiresource.py", line 153, in create response, , api_key = requestor.request( File "C:\Users\hinelon\AppData\Roaming\Python\Python38\site-packages\openai\api_requestor.py", line 226, in request resp, got_stream = self._interpret_response(result, stream) File "C:\Users\hinelon\AppData\Roaming\Python\Python38\site-packages\openai\api_requestor.py", line 619, in _interpret_response self._interpret_response_line( File "C:\Users\hinelon\AppData\Roaming\Python\Python38\site-packages\openai\api_requestor.py", line 679, in _interpret_response_line raise self.handle_error_response( openai.error.InvalidRequestError: 3 is greater than the maximum of 2 - 'temperature'

Set the environment variable HYDRA_FULL_ERROR=1 for a complete stack trace.

[GoooDte]

大家查一下easyinstruct安装的是否是最新版的

[TsuMYF]

@GoooDte 您好，我的easyinstruct是0.0.2版本，刚刚也装了最新版本的（显示最新的仍是0.0.2）。但还是有上述报错

[TsuMYF]

easyinstruct 0.0.2 openai 0.27.4 hydra-core 1.3.2

[GoooDte]

可以先将easyinstruct的代码clone到本地采用python setup.py install的方式安装 https://github.com/zjunlp/EasyInstruct 这里面是最新的代码 pip包可能不是最新的 我们今天会更新

[TsuMYF]

您好 我发现源代码中的问题了，easyinstruct\prompts\ie_prompt.py", line 276 的openai_result = super().get_openai_result(engine, temperature, max_tokens, top_p, frequency_penalty, presence_penalty) 这部分的参数是按默认顺序传入的。 但是easyinstruct\prompts\base_prompt.py", line 31定义的get_openai_result是(self, engine = "gpt-3.5-turbo", system_message: Optional[str] = "You are a helpful assistant.", temperature: Optional[float] = 0, max_tokens: Optional[int] = 64, top_p: Optional[float] = 1.0, n: Optional[int] = 1, frequency_penalty: Optional[float] = 0.0, presence_penalty: Optional[float] = 0.0 ) 如果按默认顺，easyinstruct\prompts\ie_prompt.py", line 276的参数值，会传入easyinstruct\prompts\base_prompt.py", line 31函数中错的位置上。

只需要在easyinstruct\prompts\ie_prompt.py", line 276 的openai_result函数中指定好参数值的传入对象即可：openai_result = super().get_openai_result( engine = engine,temperature = temperature,max_tokens = max_tokens,top_p = top_p,frequency_penalty = frequency_penalty,presence_penalty = presence_penalty )

[GoooDte]

感谢你的细心发现，这个bug我们已经发现了，在0.0.2之后的版本已经修复，可以先采用本地安装的方式使用最新的easyinstruct

[TsuMYF]

@GoooDte 好的，感谢您们的debug，我之后会安装新的版本

[GoooDte]

easyinstruct新版本已发布 可以直接采用pip方式安装