Diffusion Transformer

Also known as DiT

Overview

Diffusion Transformers (DiTs) use transformer networks to turn noise into data, typically by repeatedly predicting denoising updates or following a learned flow. Work explores latent- and pixel-space models, conditioning through text, images, geometry, or other modalities, and representation-learning methods that improve generation and downstream perception. To lower the cost of training and inference, researchers use sparse attention, token reduction, caching, quantization, distillation, and hardware-aware designs. A central challenge is balancing speed, memory, and model size against fidelity, temporal or spatial consistency, and reliable performance across tasks and benchmarks.

Latest papers

CardsList

519 papers46 papers in the last four weeks, down 47% on the four weeks before. 0.3% of all new papers.

  1. Phase-Aligned RoPE for Mixed-Resolution Diffusion Transformer

    Nov 24, 2025Haoyu Wu, Jingyi Xu, Qiaomu Miao +2Rotary Positional EmbeddingsVideo Diffusion Models

  2. DiTGaze: Geometry-Conditioned Diffusion Augmentation for Few-Shot Gaze Estimation

    Nov 14, 2025Abiram Panchalingam, Indu P. Bodala, Stuart E. MiddletonGaze EstimationDiffusion-Based Data Augmentation

  3. Stochastic Interpolants via Conditional Dependent Coupling

    Sep 27, 2025Chenrui Ma, Xi Xiao, Tianyang Wang +2Generative ModelingDiffusion-Based Image Generation

  4. RefAM: Attention Magnets for Zero-Shot Referral Segmentation

    Sep 26, 2025Anna Kukleva, Enis Simsar, Alessio Tonioni +4Attention MechanismsImage Segmentation

  5. NABLA: Neighborhood Adaptive Block-Level Attention

    Jul 17, 2025Dmitrii Mikhailov, Aleksey Letunovskiy, Maria Kovaleva +6Video Diffusion ModelsDiffusion Transformer

  6. Generalizable dose denoising using diffusion transformer framework for pencil beam scanning proton therapy

    Jun 4, 2025Yuzhen Ding, Hongying Feng, Jianfeng Zhao +14Diffusion-Based DenoisingHealthcare

  7. Conditioning Residuals for Diffusion Models via Representation Feedback

    May 16, 2025Weilai Xiang, Hongyu Yang, Di Huang +1Representation LearningDiffusion Transformer

  8. FiTv2: Scalable and Improved Flexible Vision Transformer for Diffusion Model

    Oct 17, 2024ZiDong Wang, Zeyu Lu, Di Huang +4Image GenerationDiffusion-Based Image Generation