Examples

Table of Contents

Install requirements

Supervised datasets collection

Conversation dataset generation

Stage1 - Supervised instructs tuning

Arg List

Stage2 - Training reward model

Features and tricks in RM training

Experiment result

Arg List

Stage3 - Training model using prompts with RL

Arg List

Inference example - After Stage3

Attention

data

Support Model

GPT

BLOOM

OPT

LLaMA

Add your own models

Actor model

Reward model

Critic model