CADS: Unleashing the Diversity of Diffusion Models through Condition-Annealed Sampling
This work addresses a key limitation in conditional diffusion models for image generation, offering a method to improve diversity without significant quality loss, which is incremental but impactful for researchers and practitioners in generative AI.
The paper tackles the problem of limited output diversity in conditional diffusion models, especially at high guidance scales or with small datasets, by introducing a sampling strategy that anneals the conditioning signal to balance diversity and condition alignment, achieving state-of-the-art FID scores of 1.70 and 2.31 on class-conditional ImageNet generation at 256x256 and 512x512 resolutions.
While conditional diffusion models are known to have good coverage of the data distribution, they still face limitations in output diversity, particularly when sampled with a high classifier-free guidance scale for optimal image quality or when trained on small datasets. We attribute this problem to the role of the conditioning signal in inference and offer an improved sampling strategy for diffusion models that can increase generation diversity, especially at high guidance scales, with minimal loss of sample quality. Our sampling strategy anneals the conditioning signal by adding scheduled, monotonically decreasing Gaussian noise to the conditioning vector during inference to balance diversity and condition alignment. Our Condition-Annealed Diffusion Sampler (CADS) can be used with any pretrained model and sampling algorithm, and we show that it boosts the diversity of diffusion models in various conditional generation tasks. Further, using an existing pretrained diffusion model, CADS achieves a new state-of-the-art FID of 1.70 and 2.31 for class-conditional ImageNet generation at 256$\times$256 and 512$\times$512 respectively.