Fine-tuning without prompt-response data

#113

by rajlohith2 - opened Apr 8

Apr 8

Hello everyone,

I'm trying to fine-tune starcoder data to improve software quality metrics of the generated code. All the implementations of fine tuning have datasets with prompt-response format. But my data consists of just the code, metrics and its issues. How do I perform fine-tuning on this data?

Thanks,
Lohith

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment