Category: Direct Preference Optimization

• OPA-DPO: An Efficient Solution for the Hallucination Problem in Multimodal Large Models