Subspace Diffusion Generative Models
This work addresses computational bottlenecks in diffusion models for researchers and practitioners, offering a method that improves sample quality and efficiency, though it is incremental as it builds on existing state-of-the-art models.
The paper tackles the computational inefficiency of high-dimensional diffusion in score-based generative models by restricting diffusion via projections onto subspaces, achieving an FID of 2.17 on unconditional CIFAR-10 while reducing inference cost.
Score-based models generate samples by mapping noise to data (and vice versa) via a high-dimensional diffusion process. We question whether it is necessary to run this entire process at high dimensionality and incur all the inconveniences thereof. Instead, we restrict the diffusion via projections onto subspaces as the data distribution evolves toward noise. When applied to state-of-the-art models, our framework simultaneously improves sample quality -- reaching an FID of 2.17 on unconditional CIFAR-10 -- and reduces the computational cost of inference for the same number of denoising steps. Our framework is fully compatible with continuous-time diffusion and retains its flexible capabilities, including exact log-likelihoods and controllable generation. Code is available at https://github.com/bjing2016/subspace-diffusion.