DiT

The 'DiT' dataset/benchmark is used to evaluate the performance of diffusion models, specifically in the context of their generative capabilities and the effectiveness of acceleration techniques like structural pruning.

Papers using DiT (3)

MosaicDiff: Training-free Structural Pruning for Diffusion Model Acceleration Reflecting Pretraining Dynamics2025Scaling Diffusion Transformers Efficiently via $\mu$P2025Faster Diffusion: Rethinking the Role of the Encoder for Diffusion Model Inference2023 · 5 cites