Recent progress in large-scale generative models has substantially advanced video generation, yet existing methods remain constrained by a rigid inference paradigm. Bidirectional diffusion models excel at global coherence and visual fidelity but suffer from slow inference, while autoregressive models offer efficient and streaming generation at the cost of long-range consistency and exposure bias. We introduce Flex-Forcing, a unified training and inference framework that enables a video diffusion model to seamlessly operate under both bidirectional and autoregressive generation regimes. The core idea is a flexible chunking mechanism jointly defined over the temporal axis and denoising steps. This design allows the model to (1) perform flexible chunking according to different device budgets, (2) perform bidirectional inference across chunks for global structure planning, while generating frames autoregressively within each chunk for efficient and fine-grained synthesis, and (3) perform any-order, any-timestep autoregressive generation without the strict causal constraint. Extensive experiments on multiple video generation benchmarks demonstrate that Flex-Forcing achieves consistently better video quality, long-video stability than strong baselines with a rigid inference schedule, while offering faster inference.
Image inpainting is an underdetermined inverse problem, which naturally allows diverse contents to fill up the missing or corrupted regions realistically. Prevalent approaches using convolutional neural networks (CNNs) can synthesize visually pleasant contents, but CNNs suffer from limited perception fields for capturing global features. With image-level attention, transformers enable to model long-range dependencies and generate diverse contents with autoregressive modeling of pixel-sequence distributions. However, the unidirectional attention in autoregressive transformers is suboptimal as c…
Autoregressive models and their sequential factorization of the data likelihood have recently demonstrated great potential for image representation and synthesis. Nevertheless, they incorporate image context in a linear 1D order by attending only to previously synthesized image patches above or to the left. Not only is this unidirectional, sequential bias of attention unnatural for images as it disregards large parts of a scene until synthesis is almost complete. It also processes the entire image on a single scale, thus ignoring more global contextual information up to the gist of the entire …
UniTemp: Unlocking Video Generation in Any Temporal Order via Bidirectional Distillation
University of Wisconsin Madison（威斯康星大学麦迪逊分校）；Adobe Research（Adobe 研究院）；University of California Los Angeles（加利福尼亚大学洛杉矶分校）；University of California Davis（加利福尼亚大学戴维斯分校）
Glance Forcing: Making Bidirectional Diffusion Autoregressive at a Glance
Wuhan University（武汉大学）；University of Electronic Science and Technology of China（电子科技大学）；Sun Yat-sen University（中山大学）；University of Washington（华盛顿大学）；National University of Singapore（新加坡国立大学）；Central South University（中南大学）
FlexVAR: Flexible Visual Autoregressive Modeling without Residual Prediction
Institute of Information Science, Beijing Jiaotong University（北京交通大学信息科学学院）；University of Technology Sydney（悉尼科技大学）；Meituan（美团）；Georgia Institute of Technology（佐治亚理工学院）
Resurrect Mask AutoRegressive Modeling for Efficient and Scalable Image Generation
Nanjing University（南京大学）；Shanghai AI Laboratory（上海人工智能实验室）；Shanghai Innovation Institute（上海创新研究院）；The Chinese University of Hong Kong（香港中文大学）；Shanghai Jiao Tong University（上海交通大学）；University of Minnesota Twin Cities（明尼苏达大学 Twin Cities 分校）
VARPose: Flexible 2D Pose Densification via Visual Autoregressive Modeling for Enhanced 3D Lifting
School of Artificial Intelligence, Sun Yat-sen University（中山大学人工智能学院）；The Technology Innovation Center for Collaborative Applications of Natural Resources Data in GBA, MNR（自然资源部粤港澳大湾区自然资源数据协同应用技术创新中心）
Autoregressive Video Generation without Vector Quantization
Beijing University of Posts and Telecommunications（北京邮电大学）；Key Laboratory of Intelligent Information Processing, ICT, CAS（中国科学院计算技术研究所智能信息处理重点实验室）；University of Chinese Academy of Sciences（中国科学院大学）；Dalian University of Technology（大连理工大学）；Beijing Academy of Artificial Intelligence（北京智源人工智能研究院）
Conventional methods for the image-text generation tasks mainly tackle the naturally bidirectional generation tasks separately, focusing on designing task-specific frameworks to improve the quality and fidelity of the generated samples. Recently, Vision-Language Pre-training models have greatly improved the performance of the image-to-text generation tasks, but large-scale pre-training models for text-to-image synthesis task are still under-developed. In this paper, we propose ERNIE-ViLG, a unified generative pre-training framework for bidirectional image-text generation with transformer model…
HER2 Expression Prediction with Flexible Multi-Modal Inputs via Dynamic Bidirectional Reconstruction
School of Computer Science and Technology, University of the Chinese Academy of Sciences（中国科学院大学计算机科学与技术学院）；School of Computer Science and Technology, University of Chinese Academy of Sciences（中国科学院大学计算机科学与技术学院）；Institute for Clarity in Documentation（清晰文档研究所）；Inria Paris-Rocquencourt（巴黎- Rocquencourt 国家信息与自动化所）；Rajiv Gandhi University（拉贾·甘地大学）；Tsinghua University（清华大学）；Palmer Research Laboratories（帕勒实验室）
Yonsei University（延世大学）；Korea Institute of Science and Technology（韩国科学技术院）；LG AI Research（LG人工智能研究）；University of Michigan（密歇根大学）；Seoul National University（首尔国立大学）
EarthMapper: Visual Autoregressive Models for Controllable Bidirectional Satellite-Map Translation
School of Electronics and Information Engineering, Harbin Institute of Technology（电子与信息工程学院，哈尔滨工业大学）；Faculty of Computing, Harbin Institute of Technology（计算机学院，哈尔滨工业大学）；Peng Cheng Lab（鹏城实验室）
Google（谷歌）；Carnegie Mellon University（卡内基梅隆大学）；Georgia Institute of Technology（佐治亚理工学院）
Creating visual layouts is a critical step in graphic design. Automatic generation of such layouts is essential for scalable and diverse visual designs. To advance conditional layout generation, we introduce BLT, a bidirectional layout transformer. BLT differs from previous work on transformers in adopting non-autoregressive transformers. In training, BLT learns to predict the masked attributes by attending to surrounding attributes in two directions. During inference, BLT first generates a draft layout from the input and then iteratively refines it into a high-quality layout by masking out lo…
