Examples

Install requirements

Train the reward model (Stage 2)

Train with dummy prompt data (Stage 3)

Train with real prompt data (Stage 3)

Inference example(After Stage3)

data

Support Model

GPT

BLOOM

OPT