How to install providers correctly?

[BokDan93]

I have a function where I alternately access different models. I added all the providers that I found in the models file for the models I needed. What if I don’t specify providers, then on what basis do we use providers to get an answer? Are they taken randomly?

What is the difference between these functions?
client = G4fClient(provider=RetryProvider([You, ChatgptNext, OpenaiChat, Bing, Liaobots]))
or
client = G4fClient()

Here is my code:

models = ["gpt-4-turbo", "gpt-3.5-turbo", "gpt-4", "gpt-3.5-turbo"]
models = models * 3

for attempt in range(1, 13):
    client = G4fClient(provider=RetryProvider([You, ChatgptNext, OpenaiChat, Bing, Liaobots]))
    loop = asyncio.get_event_loop()
    try:
        model = models[attempt - 1]
        response = await loop.run_in_executor(None, lambda: client.chat.completions.create(
            model=model,
            messages=[{"role": "user", "content": full_request}],
            web_search=False
        ))
        response_str = response.choices[0].message.content       

[hlohaus]

Hey,

when you don't choose a provider, it uses the "best_provider" from the models file. We're using the RetryProvider to pick a provider from a list of "best" providers.

[hlohaus]

You can also use the AsyncClient. Then you don't need to loop the call. It's already asynchronous. Check out the docs here: g4f docs

[BokDan93]

Thank you, I did as you said, I received an answer, but now I have an error:
Python\Python310\lib\site-packages\curl_cffi\aio.py:205: UserWarning: Curlm alread closed! quitting from process_data
warnings.warn("Curlm alread closed! quitting from process_data")