max_tokens's default in docstring of InferenceClient::chat_completion's is 20 but that of TGI is 100

[sadra-barikbin]

Hi there!🤗

InferenceClient::chat_completion's docstring states that max_tokens's default value is 20 but it seems that of the TGI is 100.

huggingface_hub/src/huggingface_hub/inference/_client.py

Lines 586 to 587 in 8a99deb

	max_tokens (`int`, *optional*):
	Maximum number of tokens allowed in the response. Defaults to 20.

https://github.com/huggingface/text-generation-inference/blob/6e3220529df5906ae586031873b7865e9923040b/router/src/lib.rs#L939

[hanouticelina]

Hello @sadra-barikbin , thanks a lot for reporting this! 🤗 I just opened a PR #2653 to fix the docstrings.