Universal Inverse Distillation for Matching Models with Real-Data Supervision (No GANs)
This addresses the computational bottleneck for users of generative models by providing a more efficient and broadly applicable distillation method, though it appears incremental as it builds on prior distillation techniques.
The paper tackles the slow inference problem of diffusion, flow, and other matching models by proposing RealUID, a universal distillation framework that incorporates real data without GANs, achieving efficient one-step generation across various model types.
While achieving exceptional generative quality, modern diffusion, flow, and other matching models suffer from slow inference, as they require many steps of iterative generation. Recent distillation methods address this by training efficient one-step generators under the guidance of a pre-trained teacher model. However, these methods are often constrained to only one specific framework, e.g., only to diffusion or only to flow models. Furthermore, these methods are naturally data-free, and to benefit from the usage of real data, it is required to use an additional complex adversarial training with an extra discriminator model. In this paper, we present RealUID, a universal distillation framework for all matching models that seamlessly incorporates real data into the distillation procedure without GANs. Our RealUID approach offers a simple theoretical foundation that covers previous distillation methods for Flow Matching and Diffusion models, and is also extended to their modifications, such as Bridge Matching and Stochastic Interpolants.