dataset commit

[parth166]

Script to create a dataset and the uploaded dataset.

1: create_sft_data.py file is a general file to create datasets based on moral stories.

• The file structure is unclean, but can be cleaned later as its a standalone file piece. Wasn't certain of the directory structure we will follow so left it as is.

2: Dependency (need a gpt config file which I haven't shared but will define it here and can be updated later in readme) gpt_config.json: { "api_key": "", "api_base": "https://api.openai.com/v1/chat/completions", "model": "gpt-3.5-turbo-0125", "max_tokens": 500, "temperature": 1.0 }

[nightlessbaron]

LGTM! Subsequent TODO added in #12