Reward Ranked Fine Tuning

[TerryMelody]

Dear authors, is there any implementation of Reward Ranked Fine Tuning you mentioned in the paper?

[Junjie-Chu]

The same question here.

It seems the repository only contains a basic implementation of zero-shot prompt, which is usually introduced in some posts/blogs.
But the core contribution of finetuning/reward model is missing.
Appreciate it if dear authors could share the core content.
They will definitely contribute to the community a lot!

Many thanks for considering my request 😊