HF RL Explorer

Feature: Support RL Training of Agent Workflows with Trinity-RFT

Feature: Support RL Training of Agent Workflows with Trinity-RFT: a task in LegoFlow-SWE (Harbor dataset). We need to add a training module that allows users to train agent workflows using reinforcement learning through Trinity-RFT. Currently, there is no built-in way to convert an existing agent…

The task

We need to add a training module that allows users to train agent workflows using reinforcement learning through Trinity-RFT. Currently, there is no built-in way to convert an existing agent workflow into a trainable format and then run distributed RL training.

Part of Lego-X/LegoFlow-SWE.