FineTuned model does not work correctly

Jai 0 Reputation points Microsoft Employee
2023-04-26T23:14:34.2733333+00:00

I have finetuned my model using (Ada/babbage/curie) with prompt and completions. But none of the fine tuned model give desired result when asked for a question/interaction. What I am doing wrong here?

Note: Davinci is not available as of now to select it as parent.

Azure OpenAI Service
Azure OpenAI Service
An Azure service that provides access to OpenAI’s GPT-3 models with enterprise capabilities.
2,945 questions
0 comments

1 answer

Sort by: Most helpful
Most helpful Newest Oldest
  1. romungi-MSFT 45,036 Reputation points Microsoft Employee
    2023-04-27T09:32:53.02+00:00

    @Jai Have you used a corresponding validation dataset along with the fine tuning job? It's possible that the training data may not be diverse enough or may not have enough examples to cover the range of questions and interactions you're looking to generate. Additionally, it's important to ensure that the training data is properly formatted and encoded in UTF-8 with a byte-order mark (BOM) and is less than 200 MB in size. Did the result data of the customized model also showcase low accuracy?

    0 comments

Your answer

Answers can be marked as Accepted Answers by the question author, which helps users to know the answer solved the author's problem.