Using Human Feedback to Fine-tune Diffusion Models without Any Reward Model

TL;DR

This paper introduces D3PO, a novel method for directly fine-tuning diffusion models using human feedback without requiring a separate reward model, reducing costs and computational resources while maintaining high-quality outputs.

Contribution

The paper proposes D3PO, a new approach that bypasses reward model training for diffusion models, enabling more efficient and cost-effective fine-tuning guided by human preferences.

Findings

D3PO achieves comparable results to reward-based methods.

The method reduces image distortion and enhances safety.

It requires less computational overhead and no reward model training.

Abstract

Using reinforcement learning with human feedback (RLHF) has shown significant promise in fine-tuning diffusion models. Previous methods start by training a reward model that aligns with human preferences, then leverage RL techniques to fine-tune the underlying models. However, crafting an efficient reward model demands extensive datasets, optimal architecture, and manual hyperparameter tuning, making the process both time and cost-intensive. The direct preference optimization (DPO) method, effective in fine-tuning large language models, eliminates the necessity for a reward model. However, the extensive GPU memory requirement of the diffusion model's denoising process hinders the direct application of the DPO method. To address this issue, we introduce the Direct Preference for Denoising Diffusion Policy Optimization (D3PO) method to directly fine-tune diffusion models. The theoretical analysis demonstrates that although D3PO omits training a reward model, it effectively functions as the optimal reward model trained using human feedback data to guide the learning process. This approach requires no training of a reward model, proving to be more direct, cost-effective, and minimizing computational overhead. In experiments, our method uses the relative scale of objectives as a proxy for human preference, delivering comparable results to methods using ground-truth rewards. Moreover, D3PO demonstrates the ability to reduce image distortion rates and generate safer images, overcoming challenges lacking robust reward models. Our code is publicly available at https://github.com/yk7333/D3PO.