Diffusion Transformer

Also known as DiT

Follow this topic Ask about this topic

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

473 papers78 papers in the last four weeks, up 100% on the four weeks before. 0.5% of all new papers.

Apr 22, 2026

GeoRelight: Learning Joint Geometrical Relighting and Reconstruction with Flexible Multi-Modal Diffusion Transformers

Relighting a person from a single photo is an attractive but ill-posed task, as a 2D image ambiguously entangles 3D geometry, intrinsic appearance, and illumination. Current methods either use sequential pipelines that suffer from error accumulation, or they do not explicitly leverage 3D geometry during relighting, which limits physical consistency. Since relighting and estimation of 3D geometry are mutually beneficial tasks, we propose a unified Multi-Modal Diffusion Transformer (DiT) that jointly solves for both: GeoRelight. We make this possible through two key technical contributions: isotropic NDC-Orthographic Depth (iNOD), a distortion-free 3D representation compatible with latent diffusion models; and a strategic mixed-data training method that combines synthetic and auto-labeled real data. By solving geometry and relighting jointly, GeoRelight achieves better performance than both sequential models and previous systems that ignored geometry.

Yuxuan Xue, Ruofan Liang, Egor Zakharov +6

Codec Avatars Lab, Meta · University of Tübingen · Max Planck Institute for Informatics, Saarland Informatics Campus

Image-to-Image Translation Diffusion Transformer Monocular 3D Reconstruction Image Relighting Multimodal Diffusion Models

Apr 20, 2026

AdaCluster: Adaptive Query-Key Clustering for Sparse Attention in Video Generation

Video diffusion transformers (DiTs) suffer from prohibitive inference latency due to quadratic attention complexity. Existing sparse attention methods either overlook semantic similarity or fail to adapt to heterogeneous token distributions across layers, leading to model performance degradation. We propose AdaCluster, a training-free adaptive clustering framework that accelerates the generation of DiTs while preserving accuracy. AdaCluster applies an angle-similarity-preserving clustering method to query vectors for higher compression, and designs a euclidean-similarity-preserving clustering method for keys, covering cluster number assignment, threshold-wise adaptive clustering, and efficient critical cluster selection. Experiments on CogVideoX-2B, HunyuanVideo, and Wan-2.1 on one A40 GPU demonstrate up to 1.67-4.31x speedup with negligible quality degradation.

Haoyue Tan, Shengnan Wang, Yulin Qiao +5

University of Science and Technology of China · Institute of Artificial Intelligence, Hefei Comprehensive National Science Center · 3Independent Researcher +2

Video Diffusion Models Diffusion Transformer Sparse Attention Diffusion Model Inference Acceleration

Apr 20, 2026

LoRaQ: Optimized Low Rank Approximation for 4-bit Quantization

Post-training quantization (PTQ) is essential for deploying large diffusion transformers on resource-constrained hardware, but aggressive 4-bit quantization significantly degrades generative performance. Low-rank approximation methods have emerged as a promising solution by appending auxiliary linear branches to restore performance. However, current state-of-the-art approaches assume these branches must retain high precision (W16A16) and rely on heavy, data-dependent calibration for initialization. We challenge both limitations with LoRaQ (Low-Rank Approximated Quantization), a simple, data-free calibration approach that optimizes quantization error compensation. By overcoming the need for high-precision branches, LoRaQ enables the first fully sub-16 bit pipeline, allowing the low-rank branch itself to be quantized. We demonstrate that, at equal memory overhead, LoRaQ outperforms the state-of-the-art methods in their native implementations on Pixart-

Σ

and SANA. We also analyze mixed-precision configurations, showing that setups such as W8A8, W6A6, and W4A8 for the low-rank branch, alongside a W4 main layer, yield superior results while maintaining a fully quantized architecture compatible with modern mixed-precision hardware.

Yann Bouquet, Alireza Khodamoradi, Sophie Yáng Shen +2

Research and Advancement Development Team, Advanced Micro Devices, Inc, USA · Computer Vision Laboratory, Ecole Polytechnique F´ed´erale de Lausanne, Lausanne, Switzerland.

Mixed-Precision Quantization Diffusion Transformer Diffusion Model Quantization 4-Bit Quantization Low-Rank Approximation Post-Training Quantization+1

Apr 20, 2026

UniCSG: Unified High-Fidelity Content-Constrained Style-Driven Generation via Staged Semantic and Frequency Disentanglement

Style transfer must match a target style while preserving content semantics. DiT-based diffusion models often suffer from content-style entanglement, leading to reference-content leakage and unstable generation. We present UniCSG, a unified framework for content-constrained, style-driven generation in both text-guided and reference-guided settings. UniCSG employs staged training: (i) a latent-space semantic disentanglement stage that combines low-frequency preprocessing with conditioning corruption to encourage content-style separation, and (ii) a latent-space frequency-aware detail reconstruction stage that refines details via multi-scale frequency supervision. We further incorporate pixel-space reward learning to align latent objectives with perceptual quality after decoding. Experiments demonstrate improved content faithfulness, style alignment, and robustness in both settings.

Jingwei Yang, Ruoxi Wu, Wei Shen +4

China University of Mining and Technology, Beijing, China · OPPO Artificial Intelligence Center, Beijing, China

Constrained Generative Modeling Diffusion Transformer Image Style Transfer Representation Disentanglement Diffusion-Based Image Generation Controllable Image Generation

Apr 17, 2026

Hierarchical Codec Diffusion for Video-to-Speech Generation

Video-to-Speech (VTS) generation aims to synthesize speech from a silent video without auditory signals. However, existing VTS methods disregard the hierarchical nature of speech, which spans coarse speaker-aware semantics to fine-grained prosodic details. This oversight hinders direct alignment between visual and speech features at specific hierarchical levels during property matching. In this paper, leveraging the hierarchical structure of Residual Vector Quantization (RVQ)-based codec, we propose HiCoDiT, a novel Hierarchical Codec Diffusion Transformer that exploits the inherent hierarchy of discrete speech tokens to achieve strong audio-visual alignment. Specifically, since lower-level tokens encode coarse speaker-aware semantics and higher-level tokens capture fine-grained prosody, HiCoDiT employs low-level and high-level blocks to generate tokens at different levels. The low-level blocks condition on lip-synchronized motion and facial identity to capture speaker-aware content, while the high-level blocks use facial expression to modulate prosodic dynamics. Finally, to enable more effective coarse-to-fine conditioning, we propose a dual-scale adaptive instance layer normalization that jointly captures global vocal style through channel-wise normalization and local prosody dynamics through temporal-wise normalization. Extensive experiments demonstrate that HiCoDiT outperforms baselines in fidelity and expressiveness, highlighting the potential of discrete modelling for VTS. The code and speech demo are both available at https://github.com/Jiaxin-Ye/HiCoDiT.

Jiaxin Ye, Gaoxiang Cong, Chenhui Wang +4

Fudan University · Institute of Computing Technology, Chinese Academy of Sciences · University of Chinese Academy of Sciences +1

Audio-Video Generation Diffusion Transformer Audio-Visual Alignment

Apr 15, 2026

OneHOI: Unifying Human-Object Interaction Generation and Editing

Human-Object Interaction (HOI) modelling captures how humans act upon and relate to objects, typically expressed as <person, action, object> triplets. Existing approaches split into two disjoint families: HOI generation synthesises scenes from structured triplets and layout, but fails to integrate mixed conditions like HOI and object-only entities; and HOI editing modifies interactions via text, yet struggles to decouple pose from physical contact and scale to multiple interactions. We introduce OneHOI, a unified diffusion transformer framework that consolidates HOI generation and editing into a single conditional denoising process driven by shared structured interaction representations. At its core, the Relational Diffusion Transformer (R-DiT) models verb-mediated relations through role- and instance-aware HOI tokens, layout-based spatial Action Grounding, a Structured HOI Attention to enforce interaction topology, and HOI RoPE to disentangle multi-HOI scenes. Trained jointly with modality dropout on our HOI-Edit-44K, along with HOI and object-centric datasets, OneHOI supports layout-guided, layout-free, arbitrary-mask, and mixed-condition control, achieving state-of-the-art results across both HOI generation and editing. Code is available at https://jiuntian.github.io/OneHOI/.

Jiun Tian Hoe, Weipeng Hu, Xudong Jiang +2

Nanyang Technological University · Sun Yat-sen University · VinUniversity +1

HOI Generation Diffusion Transformer

Apr 15, 2026

ASTRA: Enhancing Multi-Subject Generation with Retrieval-Augmented Pose Guidance and Disentangled Position Embedding

Subject-driven image generation has shown great success in creating personalized content, but its capabilities are largely confined to single subjects in common poses. Current approaches face a fundamental conflict when handling multiple subjects with complex, distinct actions: preserving individual identities while enforcing precise pose structures. This challenge often leads to identity fusion and pose distortion, as appearance and structure signals become entangled within the model's architecture. To resolve this conflict, we introduce ASTRA(Adaptive Synthesis through Targeted Retrieval Augmentation), a novel framework that architecturally disentangles subject appearance from pose structure within a unified Diffusion Transformer. ASTRA achieves this through a dual-pronged strategy. It first employs a Retrieval-Augmented Pose (RAG-Pose) pipeline to provide a clean, explicit structural prior from a curated database. Then, its core generative model learns to process these dual visual conditions using our Enhanced Universal Rotary Position Embedding (EURoPE), an asymmetric encoding mechanism that decouples identity tokens from spatial locations while binding pose tokens to the canvas. Concurrently, a Disentangled Semantic Modulation (DSM) adapter offloads the identity preservation task into the text conditioning stream. Extensive experiments demonstrate that our integrated approach achieves superior disentanglement. On our designed COCO-based complex pose benchmark, ASTRA achieves a new state-of-the-art in pose adherence, while maintaining high identity fidelity and text alignment in DreamBench.

Tianze Xia, Zijian Ning, Zonglin Zhao +1

Huazhong University of Science and Technology

Multi-Subject Image Generation Image Generation Subject-Driven Image Generation Compositional Image Generation Diffusion Transformer

Apr 15, 2026

ADP-DiT: Text-Guided Diffusion Transformer for Brain Image Generation in Alzheimer's Disease Progression

Alzheimer's disease (AD) progresses heterogeneously across individuals, motivating subject-specific synthesis of follow-up magnetic resonance imaging (MRI) to support progression assessment. While Diffusion Transformers (DiT), an emerging transformer-based diffusion model, offer a scalable backbone for image synthesis, longitudinal AD MRI generation with clinically interpretable control over follow-up time and participant metadata remains underexplored. We present ADP-DiT, an interval-aware, clinically text-conditioned diffusion transformer for longitudinal AD MRI synthesis. ADP-DiT encodes follow-up interval together with multi-domain demographic, diagnostic (CN/MCI/AD), and neuropsychological information as a natural-language prompt, enabling time-specific control beyond coarse diagnostic stages. To inject this conditioning effectively, we use dual text encoders-OpenCLIP for vision-language alignment and T5 for richer clinical-language understanding. Their embeddings are fused into DiT through cross-attention for fine-grained guidance and adaptive layer normalization for global modulation. We further enhance anatomical fidelity by applying rotary positional embeddings to image tokens and performing diffusion in a pre-trained SDXL-VAE latent space to enable efficient high-resolution reconstruction. On 3,321 longitudinal 3T T1-weighted scans from 712 participants (259,038 image slices), ADP-DiT achieves SSIM 0.8739 and PSNR 29.32 dB, improving over a DiT baseline by +0.1087 SSIM and +6.08 dB PSNR while capturing progression-related changes such as ventricular enlargement and shrinking hippocampus. These results suggest that integrating comprehensive, subject-specific clinical conditions with architectures can improve longitudinal AD MRI synthesis.

Juneyong Lee, Geonwoo Baek, Ikbeom Jang

Department of Computer Science and Engineering, Hankuk University of Foreign Studies, Seoul, Republic of Korea

Medical Image Synthesis Conditional Diffusion Models Alzheimer's Disease Diffusion Transformer

Apr 14, 2026

Brain-DiT: A Universal Multi-state fMRI Foundation Model with Metadata-Conditioned Pretraining

Current fMRI foundation models primarily rely on a limited range of brain states and mismatched pretraining tasks, restricting their ability to learn generalized representations across diverse brain states. We present Brain-DiT, a universal multi-state fMRI foundation model pretrained on 349,898 sessions from 24 datasets spanning resting, task, naturalistic, disease, and sleep states. Unlike prior fMRI foundation models that rely on masked reconstruction in the raw-signal space or a latent space, Brain-DiT adopts metadata-conditioned diffusion pretraining with a Diffusion Transformer (DiT), enabling the model to learn multi-scale representations that capture both fine-grained functional structure and global semantics. Across extensive evaluations and ablations on 7 downstream tasks, we find consistent evidence that diffusion-based generative pretraining is a stronger proxy than reconstruction or alignment, with metadata-conditioned pretraining further improving downstream performance by disentangling intrinsic neural dynamics from population-level variability. We also observe that downstream tasks exhibit distinct preferences for representational scale: ADNI classification benefits more from global semantic representations, whereas age/sex prediction comparatively relies more on fine-grained local structure.

Junfeng Xia, Wenhao Ye, Xuanye Pan +3

Department of Biomedical Engineering, Southern University of Science and Technology, China · School of Biomedical Engineering, Shenzhen University, China

Medical Imaging Foundation Models Diffusion Transformer Self-Supervised Pre-Training fMRI Foundation Models Neuroimaging Conditional Diffusion Models+1

Apr 14, 2026

FreqFormer: Hierarchical Frequency-Domain Attention with Adaptive Spectral Routing for Long-Sequence Video Diffusion Transformers

Long-sequence video diffusion transformers hit a quadratic self-attention cost that dominates runtime and memory for very long token sequences. Most efficient attention methods use one approximation everywhere, yet video features are spectrally structured: low frequencies carry global layout and coarse motion; high frequencies carry texture and fine detail. We present FreqFormer, a frequency-aware heterogeneous attention framework. Token features are split into spectral bands with different operators: dense global attention on compressed low-frequency content, structured block-sparse attention on mid frequencies, and sliding-window local attention on high frequencies. A lightweight spectral routing network allocates heads across bands using layer statistics and the diffusion timestep, shifting compute toward global structure early in denoising and detail later. Cross-band summary tokens provide cheap residual exchange. FreqFormer is paired with a fused GPU execution plan that co-schedules dense, sparse, and local branches to cut kernel launches and memory traffic. We give a consistent complexity model, an orthonormal-decomposition view of approximation, and simulation-based systems numbers (throughput, arithmetic intensity, memory traffic, duration scaling). In simulations from 64K to 1M tokens, FreqFormer substantially reduces estimated attention FLOPs and KV-related memory traffic versus dense attention while keeping a hardware-friendly pattern, supporting spectrally structured heterogeneous attention as a practical direction for long-video diffusion transformers.

Haopeng Jin

Beijing University of Posts and Telecommunications

Diffusion Model Inference Acceleration Diffusion Transformer Video Diffusion Models

Apr 14, 2026

CoSyncDiT: Cognitive Synchronous Diffusion Transformer for Movie Dubbing

Movie dubbing aims to synthesize speech that preserves the vocal identity of a reference audio while synchronizing with the lip movements in a target video. Existing methods fail to achieve precise lip-sync and lack naturalness due to explicit alignment at the duration level. While implicit alignment solutions have emerged, they remain susceptible to interference from the reference audio, triggering timbre and pronunciation degradation in in-the-wild scenarios. In this paper, we propose a novel flow matching-based movie dubbing framework driven by the Cognitive Synchronous Diffusion Transformer (CoSync-DiT), inspired by the cognitive process of professional actors. This architecture progressively guides the noise-to-speech generative trajectory by executing acoustic style adapting, fine-grained visual calibrating, and time-aware context aligning. Furthermore, we design the Joint Semantic and Alignment Regularization (JSAR) mechanism to simultaneously constrain frame-level temporal consistency on the contextual outputs and semantic consistency on the flow hidden states, ensuring robust alignment. Extensive experiments on both standard benchmarks and challenging in-the-wild dubbing benchmarks demonstrate that our method achieves the state-of-the-art performance across multiple metrics.

Gaoxiang Cong, Liang Li, Jiaxin Ye +4

Institute of Computing Technology, Chinese Academy of Sciences, Beijing, China · University of Chinese Academy of Sciences, Beijing, China · Fudan University, Shanghai, China +2

Audio-Visual Alignment Diffusion Transformer

Apr 14, 2026

Nucleus-Image: Sparse MoE for Image Generation

We present Nucleus-Image, a text-to-image generation model that establishes a new Pareto frontier in quality-versus-efficiency by matching or exceeding leading models on GenEval, DPG-Bench, and OneIG-Bench while activating only approximately 2B parameters per forward pass. Nucleus-Image employs a sparse mixture-of-experts (MoE) diffusion transformer architecture with Expert-Choice Routing that scales total model capacity to 17B parameters across 64 routed experts per layer. We adopt a streamlined architecture optimized for inference efficiency by excluding text tokens from the transformer backbone entirely and using joint attention that enables text KV sharing across timesteps. To improve routing stability when using timestep modulation, we introduce a decoupled routing design that separates timestep-aware expert assignment from timestep-conditioned expert computation. We construct a large-scale training corpus of 1.5B high-quality training pairs spanning 700M unique images through multi-stage filtering, deduplication, aesthetic tiering, and caption curation. Training follows a progressive resolution curriculum (256 to 512 to 1024) with multi-aspect-ratio bucketing at every stage, coupled with progressive sparsification of the expert capacity factor. We adopt the Muon optimizer and share our parameter grouping recipe tailored for diffusion models with timestep modulation. Nucleus-Image demonstrates that sparse MoE scaling is a highly effective path to high-quality image generation, reaching the performance of models with significantly larger active parameter budgets at a fraction of the inference cost. These results are achieved without post-training optimization of any kind: no reinforcement learning, no direct preference optimization, and no human preference tuning. We release the training recipe, making Nucleus-Image the first fully open-source MoE diffusion model at this quality.

Chandan Akiti, Ajay Modukuri, Murali Nandan Nagarapu +2

T2I Diffusion Models Diffusion Models Diffusion Transformer Diffusion Model Acceleration

Apr 13, 2026

LayerCache: Exploiting Layer-wise Velocity Heterogeneity for Efficient Flow Matching Inference

Flow Matching models achieve state-of-the-art image generation quality but incur substantial inference cost due to iterative denoising through large Transformer networks. We observe that different layer groups within a Transformer exhibit markedly heterogeneous velocity dynamics: shallow layers are highly stable and amenable to aggressive caching, while deep layers undergo large velocity changes that demand full computation. Existing caching methods, however, treat the entire Transformer as a monolithic unit, applying a single caching decision per timestep and thus failing to exploit this heterogeneity. Based on this finding, we propose LayerCache, a layer-aware caching framework that partitions the Transformer into layer groups and makes independent, per-group caching decisions at each denoising step. LayerCache introduces an adaptive JVP span K selection mechanism that leverages per-group stability measurements to balance estimation accuracy and computational savings. We formulate a three-dimensional scheduling problem over timesteps, layer groups, and JVP span, and solve it with a greedy budget allocation algorithm. On Qwen-Image (1024x1024, 50 steps), LayerCache achieves PSNR 37.46 dB (+5.38 dB over MeanCache), SSIM 0.9834, and LPIPS 0.0178 (a 70% reduction over MeanCache) at 1.37x speedup, dominating all prior caching methods on the quality-speed Pareto frontier.

Guandong Li

iFLYTEK

Flow Matching Flow-Based Generative Modeling Diffusion Transformer

Apr 13, 2026

Any 3D Scene is Worth 1K Tokens: 3D-Grounded Representation for Scene Generation at Scale

Scene-level 3D generation has long been dominated by 2D multi-view or video diffusion models. Typically, these 2D-based approaches perform scene generation in a 2D latent space, which introduces two fundamental issues: (i) representing 3D scenes via 2D views leads to significant representation redundancy, and (ii) latent space rooted in 2D inherently limits the spatial consistency of the generated scenes. In this paper, we propose to perform 3D scene generation directly within a high-dimensional implicit 3D latent space derived from powerful 2D foundation models (e.g., DINOv2, SigLIP2, DA3). Then we tame diffusion transformers to perform diffusion modeling directly in the 3D latent space, enabling 3D-grounded scene generation. In our experiments, we not only demonstrate the advantages of our method over 2D-based approaches in terms of efficiency and spatial consistency, but also comprehensively analyze the impact of 3D latent representation on 3D scene generation performance. Furthermore, we validate the favorable scalability of our method with respect to representation and model capacity.

Dongxu Wei, Qi Xu, Zhiqi Li +6

Westlake University · Afari Intelligent Drive · Zhejiang University

3D Scene Generation 3D Diffusion Models 3D Scene Representation 3D Representation Learning Diffusion Transformer

Apr 11, 2026

Transformers Learn the Optimal DDPM Denoiser for Multi-Token GMMs

Transformer-based diffusion models have demonstrated remarkable performance at generating high-quality samples. However, our theoretical understanding of the reasons for this success remains limited. For instance, existing models are typically trained by minimizing a denoising objective, which is equivalent to fitting the score function of the training data. However, we do not know why transformer-based models can match the score function for denoising, or why gradient-based methods converge to the optimal denoising model despite the non-convex loss landscape. To the best of our knowledge, this paper provides the first convergence analysis for training transformer-based diffusion models. More specifically, we consider the population Denoising Diffusion Probabilistic Model (DDPM) objective for denoising data that follow a multi-token Gaussian mixture distribution. We theoretically quantify the required number of tokens per data point and training iterations for the global convergence towards the Bayes optimal risk of the denoising objective, thereby achieving a desired score matching error. A deeper investigation reveals that the self-attention module of the trained transformer implements a mean denoising mechanism that enables the trained model to approximate the oracle Minimum Mean Squared Error (MMSE) estimator of the injected noise in the diffusion steps. Numerical experiments validate these findings.

Hongkang Li, Hancheng Min, Rene Vidal

Department of Electrical and Systems Engineering, University of Pennsylvania, Philadelphia, USA · Institute of Natural Sciences & School of Mathematical Sciences, Shanghai Jiao Tong University, Shanghai, China

Diffusion Models Diffusion-Based Denoising Diffusion Transformer Score Matching

Apr 10, 2026

SHIFT: Steering Hidden Intermediates in Flow Transformers

Diffusion models have become leading approaches for high-fidelity image generation. Recent DiT-based diffusion models, in particular, achieve strong prompt adherence while producing high-quality samples. We propose SHIFT, a simple but effective and lightweight framework for concept removal in DiT diffusion models via targeted manipulation of intermediate activations at inference time, inspired by activation steering in large language models. SHIFT learns steering vectors that are dynamically applied to selected layers and timesteps to suppress unwanted visual concepts while preserving the prompt's remaining content and overall image quality. Beyond suppression, the same mechanism can shift generations into a desired style domain or bias samples toward adding or changing target objects. We demonstrate that SHIFT provides effective and flexible control over DiT generation across diverse prompts and targets without time-consuming retraining.

Nina Konovalova, Andrey Kuznetsov, Aibek Alanov

FusionBrain Lab · HSE University · AXXX

Controllable Image Generation Concept Erasure Diffusion Model Unlearning Diffusion Transformer Activation Steering

Apr 9, 2026

HistDiT: A Structure-Aware Latent Conditional Diffusion Model for High-Fidelity Virtual Staining in Histopathology

Immunohistochemistry (IHC) is essential for assessing specific immune biomarkers like Human Epidermal growth-factor Receptor 2 (HER2) in breast cancer. However, the traditional protocols of obtaining IHC stains are resource-intensive, time-consuming, and prone to structural damages. Virtual staining has emerged as a scalable alternative, but it faces significant challenges in preserving fine-grained cellular structures while accurately translating biochemical expressions. Current state-of-the-art methods still rely on Generative Adversarial Networks (GANs) or standard convolutional U-Net diffusion models that often struggle with "structure and staining trade-offs". The generated samples are either structurally relevant but blurry, or texturally realistic but have artifacts that compromise their diagnostic use. In this paper, we introduce HistDiT, a novel latent conditional Diffusion Transformer (DiT) architecture that establishes a new benchmark for visual fidelity in virtual histological staining. The novelty introduced in this work is, a) the Dual-Stream Conditioning strategy that explicitly maintains a balance between spatial constraints via VAE-encoded latents and semantic phenotype guidance via UNI embeddings; b) the multi-objective loss function that contributes to sharper images with clear morphological structure; and c) the use of the Structural Correlation Metric (SCM) to focus on the core morphological structure for precise assessment of sample quality. Consequently, our model outperforms existing baselines, as demonstrated through rigorous quantitative and qualitative evaluations.

Aasim Bin Saleem, Amr Ahmed, Ardhendu Behera +5

Edge Hill University, Ormskirk, L39 4QP, United Kingdom · University of Nottingham Malaysia, 43500 Semenyih, Malaysia · University of Southampton Malaysia, 79100 Iskandar Puteri, Malaysia +1

Virtual Staining Computational Pathology Conditional Diffusion Models Diffusion Transformer

Apr 6, 2026

Beyond Few-Step Inference: Accelerating Video Diffusion Transformer Model Serving with Inter-Request Caching Reuse

Video Diffusion Transformer (DiT) models are a dominant approach for high-quality video generation but suffer from high inference cost due to iterative denoising. Existing caching approaches primarily exploit similarity within the diffusion process of a single request to skip redundant denoising steps. In this paper, we introduce Chorus, a caching approach that leverages similarity across requests to accelerate video diffusion model serving. Chorus achieves up to 45% speedup on industrial 4-step distilled models, where prior intra-request caching approaches are ineffective. Particularly, Chorus employs a three-stage caching strategy along the denoising process. Stage 1 performs full reuse of latent features from similar requests. Stage 2 exploits inter-request caching in specific latent regions during intermediate denoising steps. This stage is combined with Token-Guided Attention Amplification to improve semantic alignment between the generated video and the conditional prompts, thereby extending the applicability of full reuse to later denoising steps.

Hao Liu, Ye Huang, Chenghuan Huang +5

Sun Yat-sen University, China · Independent Researcher

Diffusion Model Caching Diffusion Model Inference Acceleration Diffusion Model Serving Video Diffusion Models Diffusion Transformer

Apr 4, 2026

DiffSparse: Accelerating Diffusion Transformers with Learned Token Sparsity

Diffusion models demonstrate outstanding performance in image generation, but their multi-step inference mechanism requires immense computational cost. Previous works accelerate inference by leveraging layer or token cache techniques to reduce computational cost. However, these methods fail to achieve superior acceleration performance in few-step diffusion transformer models due to inefficient feature caching strategies, manually designed sparsity allocation, and the practice of retaining complete forward computations in several steps in these token cache methods. To tackle these challenges, we propose a differentiable layer-wise sparsity optimization framework for diffusion transformer models, leveraging token caching to reduce token computation costs and enhance acceleration. Our method optimizes layer-wise sparsity allocation in an end-to-end manner through a learnable network combined with a dynamic programming solver. Additionally, our proposed two-stage training strategy eliminates the need for full-step processing in existing methods, further improving efficiency. We conducted extensive experiments on a range of diffusion-transformer models, including DiT-XL/2, PixArt-

α

, FLUX, and Wan2.1. Across these architectures, our method consistently improves efficiency without degrading sample quality. For example, on PixArt-

α

with 20 sampling steps, we reduce computational cost by

54\%

while achieving generation metrics that surpass those of the original model, substantially outperforming prior approaches. These results demonstrate that our method delivers large efficiency gains while often improving generation quality.

Haowei Zhu, Ji Liu, Ziqiong Liu +4

Advanced Micro Devices, Inc. · Tsinghua University · BNRist

Diffusion Model Inference Acceleration Diffusion Models Diffusion Transformer Diffusion Model Caching

Apr 3, 2026

LumaFlux: Lifting 8-Bit Worlds to HDR Reality with Physically-Guided Diffusion Transformers

The rapid adoption of HDR-capable devices has created a pressing need to convert the 8-bit Standard Dynamic Range (SDR) content into perceptually and physically accurate 10-bit High Dynamic Range (HDR). Existing inverse tone-mapping (ITM) methods often rely on fixed tone-mapping operators that struggle to generalize to real-world degradations, stylistic variations, and camera pipelines, frequently producing clipped highlights, desaturated colors, or unstable tone reproduction. We introduce LumaFlux, a first physically and perceptually guided diffusion transformer (DiT) for SDR-to-HDR reconstruction by adapting a large pretrained DiT. Our LumaFlux introduces (1) a Physically-Guided Adaptation (PGA) module that injects luminance, spatial descriptors, and frequency cues into attention through low-rank residuals; (2) a Perceptual Cross-Modulation (PCM) layer that stabilizes chroma and texture via FiLM conditioning from vision encoder features; and (3) an HDR Residual Coupler that fuses physical and perceptual signals under a timestep- and layer-adaptive modulation schedule. Finally, a lightweight Rational-Quadratic Spline decoder reconstructs smooth, interpretable tone fields for highlight and exposure expansion, enhancing the output of the VAE decoder to generate HDR. To enable robust HDR learning, we curate the first large-scale SDR-HDR training corpus. For fair and reproducible comparison, we further establish a new evaluation benchmark, comprising HDR references and corresponding expert-graded SDR versions. Across benchmarks, LumaFlux outperforms state-of-the-art baselines, achieving superior luminance reconstruction and perceptual color fidelity with minimal additional parameters.

Shreshth Saini, Hakan Gedik, Neil Birkbeck +3

The University of Texas at Austin · Google, Inc.

HDR Imaging Diffusion Models for Image Restoration Diffusion Transformer Physics-Guided Image Restoration

Apr 2, 2026

Crystalite: A Lightweight Transformer for Efficient Crystal Modeling

Generative models for crystalline materials often rely on equivariant graph neural networks, which capture geometric structure well but are costly to train and slow to sample. We present Crystalite, a lightweight diffusion Transformer for crystal modeling built around two simple inductive biases. The first is Subatomic Tokenization, a compact chemically structured atom representation that replaces high-dimensional one-hot encodings and is better suited to continuous diffusion. The second is the Geometry Enhancement Module (GEM), which injects periodic minimum-image pair geometry directly into attention through additive geometric biases. Together, these components preserve the simplicity and efficiency of a standard Transformer while making it better matched to the structure of crystalline materials. Crystalite achieves state-of-the-art results on crystal structure prediction benchmarks, and de novo generation performance, attaining the best S.U.N. discovery score among the evaluated baselines while sampling substantially faster than geometry-heavy alternatives.

Tin Hadži Veljković, Joshua Rosenthal, Ivor Lončarić +1

1UvA-Bosch Delta Lab · University of Amsterdam · 3Ruđer Bošković Institute

Diffusion Transformer Transformer Crystal Structure Prediction Crystal Structure Generation

Apr 2, 2026

SafeRoPE: Risk-specific Head-wise Embedding Rotation for Safe Generation in Rectified Flow Transformers

Recent Text-to-Image (T2I) models based on rectified-flow transformers (e.g., SD3, FLUX) achieve high generative fidelity but remain vulnerable to unsafe semantics, especially when triggered by multi-token interactions. Existing mitigation methods largely rely on fine-tuning or attention modulation for concept unlearning; however, their expensive computational overhead and design tailored to U-Net-based denoisers hinder direct adaptation to transformer-based diffusion models (e.g., MMDiT). In this paper, we conduct an in-depth analysis of the attention mechanism in MMDiT and find that unsafe semantics concentrate within interpretable, low-dimensional subspaces at head level, where a finite set of safety-critical heads is responsible for unsafe feature extraction. We further observe that perturbing the Rotary Positional Embedding (RoPE) applied to the query and key vectors can effectively modify some specific concepts in the generated images. Motivated by these insights, we propose SafeRoPE, a lightweight and fine-grained safe generation framework for MMDiT. Specifically, SafeRoPE first constructs head-wise unsafe subspaces by decomposing unsafe embeddings within safety-critical heads, and computes a Latent Risk Score (LRS) for each input vector via projection onto these subspaces. We then introduce head-wise RoPE perturbations that can suppress unsafe semantics without degrading benign content or image quality. SafeRoPE combines both head-wise LRS and RoPE perturbations to perform risk-specific head-wise rotation on query and key vector embeddings, enabling precise suppression of unsafe outputs while maintaining generation fidelity. Extensive experiments demonstrate that SafeRoPE achieves SOTA performance in balancing effective harmful content mitigation and utility preservation for safe generation of MMDiT. Codes are available at https://github.com/deng12yx/SafeRoPE.

Xiang Yang, Feifei Li, Mi Zhang +3

Fudan University, Shanghai, China · East China University of Science and Technology, Shanghai, China

Safe T2I Generation Diffusion Transformer Attention Control in Diffusion Models Rotary Positional Embeddings

Apr 1, 2026

MotionGrounder: Grounded Multi-Object Motion Transfer via Diffusion Transformer

Motion transfer enables controllable video generation by transferring temporal dynamics from a reference video to synthesize a new video conditioned on a target caption. However, existing Diffusion Transformer (DiT)-based methods are limited to single-object videos, restricting fine-grained control in real-world scenes with multiple objects. In this work, we introduce MotionGrounder, a DiT-based framework that firstly handles motion transfer with multi-object controllability. Our Flow-based Motion Signal (FMS) in MotionGrounder provides a stable motion prior for target video generation, while our Object-Caption Alignment Loss (OCAL) grounds object captions to their corresponding spatial regions. We further propose a new Object Grounding Score (OGS), which jointly evaluates (i) spatial alignment between source video objects and their generated counterparts and (ii) semantic consistency between each generated object and its target caption. Our experiments show that MotionGrounder consistently outperforms recent baselines across quantitative, qualitative, and human evaluations.

Samuel Teodoro, Yun Chen, Agus Gunawan +3

School of Electrical Engineering, Korea Advanced Institute of Science and Technology, Daejeon, South Korea · Adobe Research, San Jose, California · CMLab, Chung-Ang University, Seoul, South Korea

Controllable Video Generation Video Diffusion Models Diffusion Transformer

Mar 30, 2026

MMFace-DiT: A Dual-Stream Diffusion Transformer for High-Fidelity Multimodal Face Generation

Recent multimodal face generation models address the spatial control limitations of text-to-image diffusion models by augmenting text-based conditioning with spatial priors such as segmentation masks, sketches, or edge maps. This multimodal fusion enables controllable synthesis aligned with both high-level semantic intent and low-level structural layout. However, most existing approaches typically extend pre-trained text-to-image pipelines by appending auxiliary control modules or stitching together separate uni-modal networks. These ad hoc designs inherit architectural constraints, duplicate parameters, and often fail under conflicting modalities or mismatched latent spaces, limiting their ability to perform synergistic fusion across semantic and spatial domains. We introduce MMFace-DiT, a unified dual-stream diffusion transformer engineered for synergistic multimodal face synthesis. Its core novelty lies in a dual-stream transformer block that processes spatial (mask/sketch) and semantic (text) tokens in parallel, deeply fusing them through a shared Rotary Position-Embedded (RoPE) Attention mechanism. This design prevents modal dominance and ensures strong adherence to both text and structural priors to achieve unprecedented spatial-semantic consistency for controllable face generation. Furthermore, a novel Modality Embedder enables a single cohesive model to dynamically adapt to varying spatial conditions without retraining. MMFace-DiT achieves a 40% improvement in visual fidelity and prompt alignment over six state-of-the-art multimodal face generation models, establishing a flexible new paradigm for end-to-end controllable generative modeling. The code and dataset are available on our project page: https://vcbsl.github.io/MMFace-DiT/

Bharath Krishnamurthy, Ajita Rattani

University of North Texas, Denton, TX, USA

Multimodal Generation Controllable Image Generation Multimodal Diffusion Models Diffusion Transformer Multimodal Fusion

Mar 30, 2026

On-the-fly Repulsion in the Contextual Space for Rich Diversity in Diffusion Transformers

Modern Text-to-Image (T2I) diffusion models have achieved remarkable semantic alignment, yet they often suffer from a significant lack of variety, converging on a narrow set of visual solutions for any given prompt. This typicality bias presents a challenge for creative applications that require a wide range of generative outcomes. We identify a fundamental trade-off in current approaches to diversity: modifying model inputs requires costly optimization to incorporate feedback from the generative path. In contrast, acting on spatially-committed intermediate latents tends to disrupt the forming visual structure, leading to artifacts. In this work, we propose to apply repulsion in the Contextual Space as a novel framework for achieving rich diversity in Diffusion Transformers. By intervening in the multimodal attention channels, we apply on-the-fly repulsion during the transformer's forward pass, injecting the intervention between blocks where text conditioning is enriched with emergent image structure. This allows for redirecting the guidance trajectory after it is structurally informed but before the composition is fixed. Our results demonstrate that repulsion in the Contextual Space produces significantly richer diversity without sacrificing visual fidelity or semantic adherence. Furthermore, our method is uniquely efficient, imposing a small computational overhead while remaining effective even in modern "Turbo" and distilled models where traditional trajectory-based interventions typically fail.

Omer Dahary, Benaya Koren, Daniel Garibi +1

Tel Aviv University, Israel · Snap Research, Israel

T2I Generation Diffusion Model Guidance Diffusion Transformer Attention Control in Diffusion Models

Mar 30, 2026

EdgeDiT: Hardware-Aware Diffusion Transformers for Efficient On-Device Image Generation

Diffusion Transformers (DiT) have established a new state-of-the-art in high-fidelity image synthesis; however, their massive computational complexity and memory requirements hinder local deployment on resource-constrained edge devices. In this paper, we introduce EdgeDiT, a family of hardware-efficient generative transformers specifically engineered for mobile Neural Processing Units (NPUs), such as the Qualcomm Hexagon and Apple Neural Engine (ANE). By leveraging a hardware-aware optimization framework, we systematically identify and prune structural redundancies within the DiT backbone that are particularly taxing for mobile data-flows. Our approach yields a series of lightweight models that achieve a 20-30% reduction in parameters, a 36-46% decrease in FLOPs, and a 1.65-fold reduction in on-device latency without sacrificing the scaling advantages or the expressive capacity of the original transformer architecture. Extensive benchmarking demonstrates that EdgeDiT offers a superior Pareto-optimal trade-off between Frechet Inception Distance (FID) and inference latency compared to both optimized mobile U-Nets and vanilla DiT variants. By enabling responsive, private, and offline generative AI directly on-device, EdgeDiT provides a scalable blueprint for transitioning large-scale foundation models from high-end GPUs to the palm of the user.

Sravanth Kodavanti, Manjunath Arveti, Sowmya Vajrala +2

Samsung Research Institute Bangalore, India

Diffusion Transformer Diffusion-Based Image Generation Diffusion Model Inference Acceleration On-Device Inference Neural Network Optimization

Mar 29, 2026

OPRO: Orthogonal Panel-Relative Operators for Panel-Aware In-Context Image Generation

We introduce a parameter-efficient adaptation method for panel-aware in-context image generation with pre-trained diffusion transformers. The key idea is to compose learnable, panel-specific orthogonal operators onto the backbone's frozen positional encodings. This design provides two desirable properties: (1) isometry, which preserves the geometry of internal features, and (2) same-panel invariance, which maintains the model's pre-trained intra-panel synthesis behavior. Through controlled experiments, we demonstrate that the effectiveness of our adaptation method is not tied to a specific positional encoding design but generalizes across diverse positional encoding regimes. By enabling effective panel-relative conditioning, the proposed method consistently improves in-context image-based instructional editing pipelines, including state-of-the-art approaches.

Sanghyeon Lee, Minwoo Lee, Euijin Shin +3

Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Korea

Image Generation Parameter-Efficient Fine-Tuning Diffusion-Based Image Editing Diffusion Transformer

Mar 28, 2026

Spectral-Aware Text-to-Time Series Generation with Billion-Scale Multimodal Meteorological Data

Text-to-time-series generation is particularly important in meteorology, where natural language offers intuitive control over complex, multi-scale atmospheric dynamics. Existing approaches are constrained by the lack of large-scale, physically grounded multimodal datasets and by architectures that overlook the spectral-temporal structure of weather signals. We address these challenges with a unified framework for text-guided meteorological time-series generation. First, we introduce MeteoCap-3B, a billion-scale weather dataset paired with expert-level captions constructed via a Multi-agent Collaborative Captioning (MACC) pipeline, yielding information-dense and physically consistent annotations. Building on this dataset, we propose MTransformer, a diffusion-based model that enables precise semantic control by mapping textual descriptions into multi-band spectral priors through a Spectral Prompt Generator, which guides generation via frequency-aware attention. Extensive experiments on real-world benchmarks demonstrate state-of-the-art generation quality, accurate cross-modal alignment, strong semantic controllability, and substantial gains in downstream forecasting under data-sparse and zero-shot settings. Additional results on general time-series benchmarks indicate that the proposed framework generalizes beyond meteorology.

Shijie Zhang

School of Computer Science and Engineering, Northeastern University, Shenyang, China

Frequency-Domain Time Series Modeling Time Series Generation Multimodal Generation Diffusion Transformer

Mar 28, 2026

EFlow: Fast Few-Step Video Generator Training from Scratch via Efficient Solution Flow

Scaling video diffusion transformers is fundamentally bottlenecked by two compounding costs: the expensive quadratic complexity of attention per step, and the iterative sampling steps. In this work, we propose EFlow, an efficient few-step training framework, that tackles these bottlenecks simultaneously. To reduce sampling steps, we build on a solution-flow objective that learns a function mapping a noised state at time t to time s. Making this formulation computationally feasible and high-quality at video scale, however, demands two complementary innovations. First, we propose Gated Local-Global Attention, a token-droppable hybrid block which is efficient, expressive, and remains highly stable under aggressive random token-dropping, substantially reducing per-step compute. Second, we develop an efficient few-step training recipe. We propose Path-Drop Guided training to replace the expensive guidance target with a computationally cheap, weak path. Furthermore, we augment this with a Mean-Velocity Additivity regularizer to ensure high fidelity at extremely low step counts. Together, our EFlow enables a practical from-scratch training pipeline, achieving up to 2.5x higher training throughput over standard solution-flow, and 45.3x lower inference latency than standard iterative models with competitive performance on Kinetics and large-scale text-to-video datasets.

Dogyun Park, Yanyu Li, Sergey Tulyakov +1

Snap Inc. · Korea University, Seoul, Republic of Korea

Video Diffusion Models Diffusion Model Acceleration Diffusion Transformer Few-Step Diffusion Sampling Efficient ViTs Flow Matching

Mar 27, 2026

MPDiT: Multi-Patch Global-to-Local Transformer Architecture For Efficient Flow Matching and Diffusion Model

Transformer architectures, particularly Diffusion Transformers (DiTs), have become widely used in diffusion and flow-matching models due to their strong performance compared to convolutional UNets. However, the isotropic design of DiTs processes the same number of patchified tokens in every block, leading to relatively heavy computation during training process. In this work, we introduce a multi-patch transformer design in which early blocks operate on larger patches to capture coarse global context, while later blocks use smaller patches to refine local details. This hierarchical design could reduces computational cost by up to 50% in GFLOPs while achieving good generative performance. In addition, we also propose improved designs for time and class embeddings that accelerate training convergence. Extensive experiments on the ImageNet dataset demonstrate the effectiveness of our architectural choices. Code is released at: https://github.com/quandao10/MPDiT

Quan Dao, Dimitris Metaxas

Rutgers University

Diffusion Transformer Diffusion-Based Image Generation Diffusion Model Acceleration