Tong Shen

Team Wan, Ang Wang, Baole Ai et al.

This report presents Wan, a comprehensive and open suite of video foundation models designed to push the boundaries of video generation. Built upon the mainstream diffusion transformer paradigm, Wan achieves significant advancements in generative capabilities through a series of innovations, including our novel VAE, scalable pre-training strategies, large-scale data curation, and automated evaluation metrics. These contributions collectively enhance the model's performance and versatility. Specifically, Wan is characterized by four key features: Leading Performance: The 14B model of Wan, trained on a vast dataset comprising billions of images and videos, demonstrates the scaling laws of video generation with respect to both data and model size. It consistently outperforms the existing open-source models as well as state-of-the-art commercial solutions across multiple internal and external benchmarks, demonstrating a clear and significant performance superiority. Comprehensiveness: Wan offers two capable models, i.e., 1.3B and 14B parameters, for efficiency and effectiveness respectively. It also covers multiple downstream applications, including image-to-video, instruction-guided video editing, and personal video generation, encompassing up to eight tasks. Consumer-Grade Efficiency: The 1.3B model demonstrates exceptional resource efficiency, requiring only 8.19 GB VRAM, making it compatible with a wide range of consumer-grade GPUs. Openness: We open-source the entire series of Wan, including source code and all models, with the goal of fostering the growth of the video generation community. This openness seeks to significantly expand the creative possibilities of video production in the industry and provide academia with high-quality video foundation models. All the code and models are available at https://github.com/Wan-Video/Wan2.1.

Tong Shen, Jingai Yu, Dong Zhou et al.

Diffusion models have shown strong capabilities in generating high-quality images from text prompts. However, these models often require large-scale training data and significant computational resources to train, or suffer from heavy structure with high latency. To this end, we propose Efficient Multimodal Diffusion Transformer (E-MMDiT), an efficient and lightweight multimodal diffusion model with only 304M parameters for fast image synthesis requiring low training resources. We provide an easily reproducible baseline with competitive results. Our model for 512px generation, trained with only 25M public data in 1.5 days on a single node of 8 AMD MI300X GPUs, achieves 0.66 on GenEval and easily reaches to 0.72 with some post-training techniques such as GRPO. Our design philosophy centers on token reduction as the computational cost scales significantly with the token count. We adopt a highly compressive visual tokenizer to produce a more compact representation and propose a novel multi-path compression module for further compression of tokens. To enhance our design, we introduce Position Reinforcement, which strengthens positional information to maintain spatial coherence, and Alternating Subregion Attention (ASA), which performs attention within subregions to further reduce computational cost. In addition, we propose AdaLN-affine, an efficient lightweight module for computing modulation parameters in transformer blocks. Our code is available at https://github.com/AMD-AGI/Nitro-E and we hope E-MMDiT serves as a strong and practical baseline for future research and contributes to democratization of generative AI models.

Chaojie Mao, Chen-Wei Xie, Chongyang Zhong et al.

We present Wan-Image, a unified visual generation system explicitly engineered to paradigm-shift image generation models from casual synthesizers into professional-grade productivity tools. While contemporary diffusion models excel at aesthetic generation, they frequently encounter critical bottlenecks in rigorous design workflows that demand absolute controllability, complex typography rendering, and strict identity preservation. To address these challenges, Wan-Image features a natively unified multi-modal architecture by synergizing the cognitive capabilities of large language models with the high-fidelity pixel synthesis of diffusion transformers, which seamlessly translates highly nuanced user intents into precise visual outputs. It is fundamentally powered by large-scale multi-modal data scaling, a systematic fine-grained annotation engine, and curated reinforcement learning data to surpass basic instruction following and unlock expert-level professional capabilities. These include ultra-long complex text rendering, hyper-diverse portrait generation, palette-guided generation, multi-subject identity preservation, coherent sequential visual generation, precise multi-modal interactive editing, native alpha-channel generation, and high-efficiency 4K synthesis. Across diverse human evaluations, Wan-Image exceeds Seedream 5.0 Lite and GPT Image 1.5 in overall performance, reaching parity with Nano Banana Pro in challenging tasks. Ultimately, Wan-Image revolutionizes visual content creation across e-commerce, entertainment, education, and personal productivity, redefining the boundaries of professional visual synthesis.

Jinbo Xing, Zeyinzi Jiang, Yuxiang Tuo et al.

Recent unified models have made unprecedented progress in both understanding and generation. However, while most of them accept multi-modal inputs, they typically produce only single-modality outputs. This challenge of producing interleaved content is mainly due to training data scarcity and the difficulty of modeling long-range cross-modal context. To address this issue, we decompose interleaved generation into textual planning and visual consistency modeling, and introduce a framework consisting of a planner and a visualizer. The planner produces dense textual descriptions for visual content, while the visualizer synthesizes images accordingly. Under this guidance, we construct large-scale textual-proxy interleaved data (where visual content is represented in text) to train the planner, and curate reference-guided image data to train the visualizer. These designs give rise to Wan-Weaver, which exhibits emergent interleaved generation ability with long-range textual coherence and visual consistency. Meanwhile, the integration of diverse understanding and generation data into planner training enables Wan-Weaver to achieve robust task reasoning and generation proficiency. To assess the model's capability in interleaved generation, we further construct a benchmark that spans a wide range of use cases across multiple dimensions. Extensive experiments demonstrate that, even without access to any real interleaved data, Wan-Weaver achieves superior performance over existing methods.

Lianghua Huang, Wei Wang, Zhi-Fan Wu et al.

Recent research arXiv:2410.15027 arXiv:2410.23775 has highlighted the inherent in-context generation capabilities of pretrained diffusion transformers (DiTs), enabling them to seamlessly adapt to diverse visual tasks with minimal or no architectural modifications. These capabilities are unlocked by concatenating self-attention tokens across multiple input and target images, combined with grouped and masked generation pipelines. Building upon this foundation, we present ChatDiT, a zero-shot, general-purpose, and interactive visual generation framework that leverages pretrained diffusion transformers in their original form, requiring no additional tuning, adapters, or modifications. Users can interact with ChatDiT to create interleaved text-image articles, multi-page picture books, edit images, design IP derivatives, or develop character design settings, all through free-form natural language across one or more conversational rounds. At its core, ChatDiT employs a multi-agent system comprising three key components: an Instruction-Parsing agent that interprets user-uploaded images and instructions, a Strategy-Planning agent that devises single-step or multi-step generation actions, and an Execution agent that performs these actions using an in-context toolkit of diffusion transformers. We thoroughly evaluate ChatDiT on IDEA-Bench arXiv:2412.11767, comprising 100 real-world design tasks and 275 cases with diverse instructions and varying numbers of input and target images. Despite its simplicity and training-free approach, ChatDiT surpasses all competitors, including those specifically designed and trained on extensive multi-task datasets. We further identify key limitations of pretrained DiTs in zero-shot adapting to tasks. We release all code, agents, results, and intermediate outputs to facilitate further research at https://github.com/ali-vilab/ChatDiT

Wenfei Guan, Jilin Mei, Tong Shen et al.

Deep learning has advanced vectorized road extraction in urban settings, yet off-road environments remain underexplored and challenging. A significant domain gap causes advanced models to fail in wild terrains due to two key issues: lack of large-scale vectorized datasets and structural weakness in prevailing methods. Models such as SAM-Road employ a node-centric paradigm that reasons at sparse endpoints, making them fragile to occlusions and ambiguous junctions in off-road scenes, leading to topological errors. This work addresses these limitations in two complementary ways. First, we release WildRoad, a global off-road road network dataset constructed efficiently with a dedicated interactive annotation tool tailored for road-network labeling. Second, we introduce MaGRoad (Mask-aware Geodesic Road network extractor), a path-centric framework that aggregates multi-scale visual evidence along candidate paths to infer connectivity robustly. Extensive experiments show that MaGRoad achieves state-of-the-art performance on our challenging WildRoad benchmark while generalizing well to urban datasets. A streamlined pipeline also yields roughly 2.5x faster inference, improving practical applicability. Together, the dataset and path-centric paradigm provide a stronger foundation for mapping roads in the wild. We release both the dataset and code at https://github.com/xiaofei-guan/MaGRoad.

Haoran Wang, Tong Shen, Wei Zhang et al.

Despite great progress in supervised semantic segmentation,a large performance drop is usually observed when deploying the model in the wild. Domain adaptation methods tackle the issue by aligning the source domain and the target domain. However, most existing methods attempt to perform the alignment from a holistic view, ignoring the underlying class-level data structure in the target domain. To fully exploit the supervision in the source domain, we propose a fine-grained adversarial learning strategy for class-level feature alignment while preserving the internal structure of semantics across domains. We adopt a fine-grained domain discriminator that not only plays as a domain distinguisher, but also differentiates domains at class level. The traditional binary domain labels are also generalized to domain encodings as the supervision signal to guide the fine-grained feature alignment. An analysis with Class Center Distance (CCD) validates that our fine-grained adversarial strategy achieves better class-level alignment compared to other state-of-the-art methods. Our method is easy to implement and its effectiveness is evaluated on three classical domain adaptation tasks, i.e., GTA5 to Cityscapes, SYNTHIA to Cityscapes and Cityscapes to Cross-City. Large performance gains show that our method outperforms other global feature alignment based and class-wise alignment based counterparts. The code is publicly available at https://github.com/JDAI-CV/FADA.

Tong Shen, Dong Li, Ziheng Gao et al.

Video Frame Interpolation (VFI) is a crucial technique in various applications such as slow-motion generation, frame rate conversion, video frame restoration etc. This paper introduces an efficient video frame interpolation framework that aims to strike a favorable balance between efficiency and quality. Our framework follows a general paradigm consisting of a flow estimator and a refinement module, while incorporating carefully designed components. First of all, we adopt depth-wise convolution with large kernels in the flow estimator that simultaneously reduces the parameters and enhances the receptive field for encoding rich context and handling complex motion. Secondly, diverging from a common design for the refinement module with a UNet-structure (encoder-decoder structure), which we find redundant, our decoder-only refinement module directly enhances the result from coarse to fine features, offering a more efficient process. In addition, to address the challenge of handling high-definition frames, we also introduce an innovative HD-aware augmentation strategy during training, leading to consistent enhancement on HD images. Extensive experiments are conducted on diverse datasets, Vimeo90K, UCF101, Xiph and SNU-FILM. The results demonstrate that our approach achieves state-of-the-art performance with clear improvement while requiring much less FLOPs and parameters, reaching to a better spot for balancing efficiency and quality.

Tong Shen, Zeru Zhu, Ji Liu

This paper develops a distributed variational quantum algorithm for solving large-scale linear equations. For a linear system of the form $Ax=b$, the large square matrix $A$ is partitioned into smaller square block submatrices, each of which is known only to a single noisy intermediate-scale quantum (NISQ) computer. Each NISQ computer communicates with certain other quantum computers in the same row and column of the block partition, where the communication patterns are described by the row- and column-neighbor graphs, both of which are connected. The proposed algorithm integrates a variant of the variational quantum linear solver at each computer with distributed classical optimization techniques. The derivation of the quantum cost function provides insight into the design of the distributed algorithm. Numerical quantum simulations demonstrate that the proposed distributed quantum algorithm can solve linear systems whose size scales with the number of computers and is therefore not limited by the capacity of a single quantum computer.

Zirui Zheng, Takashi Isobe, Tong Shen et al.

While autoregressive (AR) models have demonstrated remarkable success in image generation, extending them to layout-conditioned generation remains challenging due to the sparse nature of layout conditions and the risk of feature entanglement. We present Structured Masking for AR-based Layout-to-Image (SMARLI), a novel framework for layoutto-image generation that effectively integrates spatial layout constraints into AR-based image generation. To equip AR model with layout control, a specially designed structured masking strategy is applied to attention computation to govern the interaction among the global prompt, layout, and image tokens. This design prevents mis-association between different regions and their descriptions while enabling sufficient injection of layout constraints into the generation process. To further enhance generation quality and layout accuracy, we incorporate Group Relative Policy Optimization (GRPO) based post-training scheme with specially designed layout reward functions for next-set-based AR models. Experimental results demonstrate that SMARLI is able to seamlessly integrate layout tokens with text and image tokens without compromising generation quality. It achieves superior layoutaware control while maintaining the structural simplicity and generation efficiency of AR models.

Tong Shen, Jiawei Zuo, Fan Shi et al.

We demonstrate ViDA-MAN, a digital-human agent for multi-modal interaction, which offers realtime audio-visual responses to instant speech inquiries. Compared to traditional text or voice-based system, ViDA-MAN offers human-like interactions (e.g, vivid voice, natural facial expression and body gestures). Given a speech request, the demonstration is able to response with high quality videos in sub-second latency. To deliver immersive user experience, ViDA-MAN seamlessly integrates multi-modal techniques including Acoustic Speech Recognition (ASR), multi-turn dialog, Text To Speech (TTS), talking heads video generation. Backed with large knowledge base, ViDA-MAN is able to chat with users on a number of topics including chit-chat, weather, device control, News recommendations, booking hotels, as well as answering questions via structured knowledge.

Tong Sha, Wei Zhang, Tong Shen et al.

Deep person generation has attracted extensive research attention due to its wide applications in virtual agents, video conferencing, online shopping and art/movie production. With the advancement of deep learning, visual appearances (face, pose, cloth) of a person image can be easily generated or manipulated on demand. In this survey, we first summarize the scope of person generation, and then systematically review recent progress and technical trends in deep person generation, covering three major tasks: talking-head generation (face), pose-guided person generation (pose) and garment-oriented person generation (cloth). More than two hundred papers are covered for a thorough overview, and the milestone works are highlighted to witness the major technical breakthrough. Based on these fundamental tasks, a number of applications are investigated, e.g., virtual fitting, digital human, generative data augmentation. We hope this survey could shed some light on the future prospects of deep person generation, and provide a helpful foundation for full applications towards digital human.

Jia Li, Tong Shen, Wei Zhang et al.

The stunning progress in face manipulation methods has made it possible to synthesize realistic fake face images, which poses potential threats to our society. It is urgent to have face forensics techniques to distinguish those tampered images. A large scale dataset "FaceForensics++" has provided enormous training data generated from prominent face manipulation methods to facilitate anti-fake research. However, previous works focus more on casting it as a classification problem by only considering a global prediction. Through investigation to the problem, we find that training a classification network often fails to capture high quality features, which might lead to sub-optimal solutions. In this paper, we zoom in on the problem by conducting a pixel-level analysis, i.e. formulating it as a pixel-level segmentation task. By evaluating multiple architectures on both segmentation and classification tasks, We show the superiority of viewing the problem from a segmentation perspective. Different ablation studies are also performed to investigate what makes an effective and efficient anti-fake model. Strong baselines are also established, which, we hope, could shed some light on the field of face forensics.

Fuxun Yu, Di Wang, Yinpeng Chen et al.

Current state-of-the-art object detectors can have significant performance drop when deployed in the wild due to domain gaps with training data. Unsupervised Domain Adaptation (UDA) is a promising approach to adapt models for new domains/environments without any expensive label cost. However, without ground truth labels, most prior works on UDA for object detection tasks can only perform coarse image-level and/or feature-level adaptation by using adversarial learning methods. In this work, we show that such adversarial-based methods can only reduce the domain style gap, but cannot address the domain content distribution gap that is shown to be important for object detectors. To overcome this limitation, we propose the Cross-Domain Semi-Supervised Learning (CDSSL) framework by leveraging high-quality pseudo labels to learn better representations from the target domain directly. To enable SSL for cross-domain object detection, we propose fine-grained domain transfer, progressive-confidence-based label sharpening and imbalanced sampling strategy to address two challenges: (i) non-identical distribution between source and target domain data, (ii) error amplification/accumulation due to noisy pseudo labeling on the target domain. Experiment results show that our proposed approach consistently achieves new state-of-the-art performance (2.2% - 9.5% better than prior best work on mAP) under various domain gap scenarios. The code will be released.

Tong Shen, Dong Gong, Wei Zhang et al.

Training a semantic segmentation model requires a large amount of pixel-level annotation, hampering its application at scale. With computer graphics, we can generate almost unlimited training data with precise annotation. However,a deep model trained with synthetic data usually cannot directly generalize well to realistic images due to domain shift. It has been observed that highly confident labels for the unlabeled real images may be predicted relying on the labeled synthetic data. To tackle the unsupervised domain adaptation problem, we explore the possibilities to generate high-quality labels as proxy labels to supervise the training on target data. Specifically, we propose a novel proxy-based method using multi-adversarial training. We first train the model using synthetic data (source domain). Multiple discriminators are used to align the features be-tween the source and target domain (real images) at different levels. Then we focus on obtaining and selecting high-quality proxy labels by incorporating both the confidence of the class predictor and that from the adversarial discriminators. Our discriminators not only work as a regularizer to encourage feature alignment but also provide an alternative confidence measure for generating proxy labels. Relying on the generated high-quality proxies, our model can be trained in a "supervised manner" on the target do-main. On two major tasks, GTA5->Cityscapes and SYNTHIA->Cityscapes, our method achieves state-of-the-art results, outperforming the previous by a large margin.

Xinyu Li, Wei Zhang, Tong Shen et al.

Selfie and cartoon are two popular artistic forms that are widely presented in our daily life. Despite the great progress in image translation/stylization, few techniques focus specifically on selfie cartoonization, since cartoon images usually contain artistic abstraction (e.g., large smoothing areas) and exaggeration (e.g., large/delicate eyebrows). In this paper, we address this problem by proposing a selfie cartoonization Generative Adversarial Network (scGAN), which mainly uses an attentive adversarial network (AAN) to emphasize specific facial regions and ignore low-level details. More specifically, we first design a cycle-like architecture to enable training with unpaired data. Then we design three losses from different aspects. A total variation loss is used to highlight important edges and contents in cartoon portraits. An attentive cycle loss is added to lay more emphasis on delicate facial areas such as eyes. In addition, a perceptual loss is included to eliminate artifacts and improve robustness of our method. Experimental results show that our method is capable of generating different cartoon styles and outperforms a number of state-of-the-art methods.

Tianyi Zhang, Guosheng Lin, Jianfei Cai et al.

Weakly supervised semantic segmentation receives much research attention since it alleviates the need to obtain a large amount of dense pixel-wise ground-truth annotations for the training images. Compared with other forms of weak supervision, image labels are quite efficient to obtain. In our work, we focus on the weakly supervised semantic segmentation with image label annotations. Recent progress for this task has been largely dependent on the quality of generated pseudo-annotations. In this work, inspired by spatial neural-attention for image captioning, we propose a decoupled spatial neural attention network for generating pseudo-annotations. Our decoupled attention structure could simultaneously identify the object regions and localize the discriminative parts which generates high-quality pseudo-annotations in one forward path. The generated pseudo-annotations lead to the segmentation results which achieve the state-of-the-art in weakly-supervised semantic segmentation.

Tong Shen, Guosheng Lin, Lingqiao Liu et al.

Training a Fully Convolutional Network (FCN) for semantic segmentation requires a large number of masks with pixel level labelling, which involves a large amount of human labour and time for annotation. In contrast, web images and their image-level labels are much easier and cheaper to obtain. In this work, we propose a novel method for weakly supervised semantic segmentation with only image-level labels. The method utilizes the internet to retrieve a large number of images and uses a large scale co-segmentation framework to generate masks for the retrieved images. We first retrieve images from search engines, e.g. Flickr and Google, using semantic class names as queries, e.g. class names in the dataset PASCAL VOC 2012. We then use high quality masks produced by co-segmentation on the retrieved images as well as the target dataset images with image level labels to train segmentation networks. We obtain an IoU score of 56.9 on test set of PASCAL VOC 2012, which reaches the state-of-the-art performance.

Tong Shen, Guosheng Lin, Chunhua Shen et al.

Semantic image segmentation is a fundamental task in image understanding. Per-pixel semantic labelling of an image benefits greatly from the ability to consider region consistency both locally and globally. However, many Fully Convolutional Network based methods do not impose such consistency, which may give rise to noisy and implausible predictions. We address this issue by proposing a dense multi-label network module that is able to encourage the region consistency at different levels. This simple but effective module can be easily integrated into any semantic segmentation systems. With comprehensive experiments, we show that the dense multi-label can successfully remove the implausible labels and clear the confusion so as to boost the performance of semantic segmentation systems.