Direct Distillation: A Novel Approach for Efficient Diffusion Model Inference.

Diffusion models are among the most common techniques used for image generation, having achieved state-of-the-art performance by implementing auto-regressive algorithms. However, multi-step inference processes are typically slow and require extensive computational resources. To address this issue, we propose the use of an information bottleneck to reschedule inference using a new sampling strategy, which employs a lightweight distilled neural network to map intermediate stages to the final output.