​
QianfanConfig

class QianfanConfig(BaseConfig):

Defines the parameters for generating chat completions using OpenAI compatibility.

Reference: https://cloud.baidu.com/doc/qianfan-api/s/3m7of64lb

Parameters:

  • presence_penalty (float, optional): Number between :obj:-2.0 and :obj:2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model’s likelihood to talk about new topics. See more information about frequency and presence penalties. (default: :obj:None)
  • frequency_penalty (float, optional): Number between :obj:-2.0 and :obj:2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model’s likelihood to repeat the same line verbatim. See more information about frequency and presence penalties. (default: :obj:None)
  • repetition_penalty (float, optional): Penalizes new tokens based on their appearance in the prompt and generated text. (default: :obj:None)
  • stream (bool, optional): Whether to stream the response. (default: :obj:None)
  • temperature (float, optional): Controls randomness in the response. Higher values make output more random, lower values make it more deterministic. Range: [0.0, 2.0]. (default: :obj:None)
  • top_p (float, optional): Controls diversity via nucleus sampling.
  • Range: [0.0, 1.0]. (default: :obj:None)
  • logit_bias (dict, optional): Modify the likelihood of specified tokens appearing in the completion. Accepts a json object that maps tokens (specified by their token ID in the tokenizer) to an associated bias value from :obj:-100 to :obj:100. Mathematically, the bias is added to the logits generated by the model prior to sampling. The exact effect will vary per model, but values between:obj: -1
  • and: obj:1 should decrease or increase likelihood of selection; values like :obj:-100 or :obj:100 should result in a ban or exclusive selection of the relevant token. (default: :obj:None)
  • max_tokens (Union[int, NotGiven], optional): Maximum number of tokens to generate. If not provided, model will use its default maximum. (default: :obj:None)
  • stop (Optional[List[str]], optional): List of stop sequences. (default: :obj:None)
  • n (Optional[int], optional): Number of chat completion choices to generate for each input message. (default: :obj:None)
  • tools (List, optional): Specifies an array of tools that the model can call. It can contain one or more tool objects. During a function call process, the model will select one tool from the array. (default: :obj:None)