Skip to content

Rewards functions for Command Line Interfaces GRPO trainer #3448

@wa008

Description

@wa008

Feature request

Is this feature available now?

Motivation

train grpo by cli with custom reward functions

Your contribution

I can try program for this feature

Metadata

Metadata

Assignees

No one assigned

    Labels

    ✨ enhancementNew feature or request🏋 GRPORelated to GRPO🏋 RewardRelated to Reward modelling📱 cliRelated to the Command-line interface

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions