RUCAIBox/RLMEC

Features

• Feedback Alignment - Improves models via fine-grained reinforcement learning with editing constraints.
• Reinforcement Learning - Improves models via fine-grained reinforcement learning with editing constraints.