ACDiT: Interpolating Autoregressive Conditional Modeling and Diffusion Transformer
This work addresses the challenge of long-horizon visual generation for AI researchers, offering a novel hybrid approach that is incremental in integrating existing methods.
The paper tackles the problem of modeling continuous visual information by combining autoregressive and diffusion paradigms, resulting in ACDiT, which outperforms autoregressive baselines in image and video generation tasks under similar model scales.
We present ACDiT, a novel Autoregressive blockwise Conditional Diffusion Transformer, that innovatively combines autoregressive and diffusion paradigms for modeling continuous visual information. By introducing a block-wise autoregressive unit, ACDiT offers a flexible interpolation between token-wise autoregression and full-sequence diffusion, bypassing the limitations of discrete tokenization. The generation of each block is formulated as a conditional diffusion process, conditioned on prior blocks. ACDiT is easy to implement, as simple as creating a Skip-Causal Attention Mask (SCAM) on standard diffusion transformer during training. During inference, the process iterates between diffusion denoising and autoregressive decoding that can make full use of KV-Cache. We show that ACDiT performs best among all autoregressive baselines under similar model scales on image and video generation tasks. We also demonstrate that benefiting from autoregressive modeling, pretrained ACDiT can be transferred in visual understanding tasks despite being trained with the diffusion objective. The analysis of the trade-off between autoregressive modeling and diffusion demonstrates the potential of ACDiT to be used in long-horizon visual generation tasks. We hope that ACDiT offers a novel perspective on visual autoregressive generation and unlocks new avenues for unified models.