You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flink.apache.org by "Yunfeng Zhou (Jira)" <ji...@apache.org> on 2025/10/27 07:36:00 UTC

[jira] [Created] (FLINK-38572) Support controlling randomness and format of Chat Model Function

Yunfeng Zhou created FLINK-38572:
------------------------------------

             Summary: Support controlling randomness and format of Chat Model Function
                 Key: FLINK-38572
                 URL: https://issues.apache.org/jira/browse/FLINK-38572
             Project: Flink
          Issue Type: Improvement
          Components: Library / Machine Learning
    Affects Versions: 2.2.0
            Reporter: Yunfeng Zhou


In order to better support controlling the randomness and format of Chat Model Function, we propose to pass the following OpenAI API parameters through model function parameter.

* presence-penalty: Number between -2.0 and 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.
* n: How many chat completion choices to generate for each input message.
* seed: the random seed to generate results
* response-format: The format of the response, e.g., 'text' or 'json_object'.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)