A problem met when trying to use the gradio_client API to do the prediction

[LampTales]

I ran webui.py to do the prediction, and followed the instructions on the generated web-page to test the Python API.
The instructions on the page are shown here.

Then I copied the code and got the error report:

File "C:\Users\ASUS\Desktop\materials\Try\YuxiaLin\src\test_server.py", line 63, in generate
    result = self.client.predict(
  File "C:\Users\ASUS\.conda\envs\pytorch\lib\site-packages\gradio_client\client.py", line 389, in predict
    return self.submit(*args, api_name=api_name, fn_index=fn_index).result()
  File "C:\Users\ASUS\.conda\envs\pytorch\lib\site-packages\gradio_client\client.py", line 1590, in result
    return super().result(timeout=timeout)
  File "C:\Users\ASUS\.conda\envs\pytorch\lib\concurrent\futures\_base.py", line 446, in result
    return self.__get_result()
  File "C:\Users\ASUS\.conda\envs\pytorch\lib\concurrent\futures\_base.py", line 391, in __get_result
    raise self._exception
  File "C:\Users\ASUS\.conda\envs\pytorch\lib\concurrent\futures\thread.py", line 58, in run
    result = self.fn(*self.args, **self.kwargs)
  File "C:\Users\ASUS\.conda\envs\pytorch\lib\site-packages\gradio_client\client.py", line 1258, in _inner
    predictions = _predict(*data)
  File "C:\Users\ASUS\.conda\envs\pytorch\lib\site-packages\gradio_client\client.py", line 1309, in _predict
    raise KeyError(
KeyError: 'Could not find \'data\' key in response. Response received: {\'detail\': [{\'type\': \'model_attributes_type\', \'loc\': [\'body\'], \'msg\': \'Input should be a valid dictionary or object to extract fields from\', \'input\': \'{"data": ["Textbox component", "lin_en", 0.5, 0.6, 0.9, 1, "EN", null, "Happy", "Text prompt", "", 0.7], "fn_index": 0, "session_hash": "cb994f88-6334-4653-a51e-2df505a94f28"}\', \'url\': \'https://errors.pydantic.dev/2.6/v/model_attributes_type\'}]}'

My code goes like:

result = self.client.predict(
		    "hello?",	# str  in '输入文本内容' Textbox component
		    "lin_en,lin_en",	# str (Option from: [('lin_jp', 'lin_jp'), ('lin_en', 'lin_en'), ('lin_zh', 'lin_zh')]) in 'Speaker' Dropdown component
		    0.5,	# int | float (numeric value between 0 and 1) in 'SDP Ratio' Slider component
		    0.6,	# int | float (numeric value between 0.1 and 2) in 'Noise' Slider component
		    0.9,	# int | float (numeric value between 0.1 and 2) in 'Noise_W' Slider component
		    1,	# int | float (numeric value between 0.1 and 2) in 'Length' Slider component
		    "EN,EN",	# str (Option from: [('ZH', 'ZH'), ('JP', 'JP'), ('EN', 'EN'), ('mix', 'mix'), ('auto', 'auto')]) in 'Language' Dropdown component
		    None,	# str (filepath on your computer (or URL) of file) in 'Audio prompt' Audio component
		    "Happy",	# str  in 'Text prompt' Textbox component
		    "Text prompt",	# str  in 'Prompt Mode' Radio component
		    "",	# str  in '辅助文本' Textbox component
		    0.7,	# int | float (numeric value between 0 and 1) in 'Weight' Slider component
		    fn_index=0
        )
        print(result)

Then I checked the net package sent from the web-page, here is the predict package:

Thus I tried turning "lin_en,lin_en" to "lin_en", "EN,EN" to "EN", but I still got the same error report.

I think the problem is that the params are passed in a wrong format, but I have no idea what should I do to correct it.
Or are there other methods I can do the prediction via Python API?
Thank you very much if you have some suggestions.

[Stardust-minus]

You should try server_fastapi.py if you want to use bv2 via api.

[LampTales]

I have been only focusing on the master branch, so I didn't realize that there exists such a .py file.
Thank you for helping, I will try it out.😺