Han Liang

Team Seedance, De Chen, Liyang Chen et al. · gatech

Seedance 2.0 is a new native multi-modal audio-video generation model, officially released in China in early February 2026. Compared with its predecessors, Seedance 1.0 and 1.5 Pro, Seedance 2.0 adopts a unified, highly efficient, and large-scale architecture for multi-modal audio-video joint generation. This allows it to support four input modalities: text, image, audio, and video, by integrating one of the most comprehensive suites of multi-modal content reference and editing capabilities available in the industry to date. It delivers substantial, well-rounded improvements across all key sub-dimensions of video and audio generation. In both expert evaluations and public user tests, the model has demonstrated performance on par with the leading levels in the field. Seedance 2.0 supports direct generation of audio-video content with durations ranging from 4 to 15 seconds, with native output resolutions of 480p and 720p. For multi-modal inputs as reference, its current open platform supports up to 3 video clips, 9 images, and 3 audio clips. In addition, we provide Seedance 2.0 Fast version, an accelerated variant of Seedance 2.0 designed to boost generation speed for low-latency scenarios. Seedance 2.0 has delivered significant improvements to its foundational generation capabilities and multi-modal generation performance, bringing an enhanced creative experience for end users.

Han Liang, Wenqian Zhang, Wenxuan Li et al.

We have recently seen tremendous progress in diffusion advances for generating realistic human motions. Yet, they largely disregard the multi-human interactions. In this paper, we present InterGen, an effective diffusion-based approach that incorporates human-to-human interactions into the motion diffusion process, which enables layman users to customize high-quality two-person interaction motions, with only text guidance. We first contribute a multimodal dataset, named InterHuman. It consists of about 107M frames for diverse two-person interactions, with accurate skeletal motions and 23,337 natural language descriptions. For the algorithm side, we carefully tailor the motion diffusion model to our two-person interaction setting. To handle the symmetry of human identities during interactions, we propose two cooperative transformer-based denoisers that explicitly share weights, with a mutual attention mechanism to further connect the two denoising processes. Then, we propose a novel representation for motion input in our interaction diffusion model, which explicitly formulates the global relations between the two performers in the world frame. We further introduce two novel regularization terms to encode spatial relations, equipped with a corresponding damping scheme during the training of our interaction diffusion model. Extensive experiments validate the effectiveness and generalizability of InterGen. Notably, it can generate more diverse and compelling two-person motions than previous methods and enables various downstream applications for human interactions.

Yiming Ren, Chengfeng Zhao, Yannan He et al.

We propose a multi-sensor fusion method for capturing challenging 3D human motions with accurate consecutive local poses and global trajectories in large-scale scenarios, only using single LiDAR and 4 IMUs, which are set up conveniently and worn lightly. Specifically, to fully utilize the global geometry information captured by LiDAR and local dynamic motions captured by IMUs, we design a two-stage pose estimator in a coarse-to-fine manner, where point clouds provide the coarse body shape and IMU measurements optimize the local actions. Furthermore, considering the translation deviation caused by the view-dependent partial point cloud, we propose a pose-guided translation corrector. It predicts the offset between captured points and the real root locations, which makes the consecutive movements and trajectories more precise and natural. Moreover, we collect a LiDAR-IMU multi-modal mocap dataset, LIPD, with diverse human actions in long-range scenarios. Extensive quantitative and qualitative experiments on LIPD and other open datasets all demonstrate the capability of our approach for compelling motion capture in large-scale scenarios, which outperforms other methods by an obvious margin. We will release our code and captured dataset to stimulate future research.

Han Liang, Yannan He, Chengfeng Zhao et al.

Monocular 3D motion capture (mocap) is beneficial to many applications. The use of a single camera, however, often fails to handle occlusions of different body parts and hence it is limited to capture relatively simple movements. We present a light-weight, hybrid mocap technique called HybridCap that augments the camera with only 4 Inertial Measurement Units (IMUs) in a learning-and-optimization framework. We first employ a weakly-supervised and hierarchical motion inference module based on cooperative Gated Recurrent Unit (GRU) blocks that serve as limb, body and root trackers as well as an inverse kinematics solver. Our network effectively narrows the search space of plausible motions via coarse-to-fine pose estimation and manages to tackle challenging movements with high efficiency. We further develop a hybrid optimization scheme that combines inertial feedback and visual cues to improve tracking accuracy. Extensive experiments on various datasets demonstrate HybridCap can robustly handle challenging movements ranging from fitness actions to Latin dance. It also achieves real-time performance up to 60 fps with state-of-the-art accuracy.

Bin Li, Ruichi Zhang, Han Liang et al.

Enabling physics-based humanoids to execute diverse behaviors from high-level textual commands remains a significant challenge. Existing methods typically follow either a two-stage paradigm that combines kinematic motion generation with physics-based tracking, or an end-to-end imitation-learning paradigm that directly generates actions from text. However, the former suffers from the inherent domain shift between kinematic generation and physics-based tracking, while the latter struggles with the substantial modality gap between textual commands and low-level actions, limiting effective semantic alignment. Notably, humanoid states encode rich motion dynamics that are more semantically aligned with textual descriptions than low-level actions, making them a natural basis for deriving behavioral intent. Building upon this insight, we propose MIND, a novel end-to-end diffusion framework for text-driven physics-based humanoid control that leverages behavioral intent as a semantic bridge between textual commands and low-level actions. At its core, MIND introduces a multi-scale intent diffusion mechanism, where a holistic intent predictor captures global behavioral dynamics to guide overall behavior synthesis, while an immediate intent predictor provides step-wise, fine-grained signals for local behavior refinement at each diffusion step. This hierarchical intent formulation imposes a structured inductive bias for humanoid control, improving semantic alignment and behavioral naturalness. Furthermore, MIND encodes humanoid states into a latent space to enable more effective semantic intent modeling. Extensive experiments demonstrate that MIND outperforms existing methods and synthesizes coherent, physically plausible, and semantically aligned humanoid behaviors from text commands. Our code will be released to facilitate future research.

Jingyan Zhang, Han Liang, Ruichi Zhang et al.

Controlling physics-based humanoids from natural-language instructions is a critical step toward general-purpose embodied agents. However, existing methods remain constrained by a tension between semantic expressiveness and physical feasibility, often failing to jointly achieve faithful instruction following, high-quality motion, and stable long-horizon control. We propose SCRIPT, a scalable diffusion policy with a multi-stage training framework for language-driven physics-based humanoid control. The core of SCRIPT is a Joint Action-State-Text Diffusion Transformer (JAST-DiT), which represents actions, physical states, and text as dedicated token streams and couples them through joint attention, enabling direct interaction between language semantics and control dynamics. To stabilize autoregressive control, we introduce a nonlinear history conditioning mechanism, which preserves the dense recent context and samples increasingly sparse cues from long-term history. Beyond supervised imitation pre-training, we propose a post-training stage, further improving the performance using Reinforcement Learning with Hybrid Rewards (RLHR). By injecting learnable noise into the flow-sampling process, RLHR effectively improves motion quality and instruction following within closed-loop simulations using hybrid physical feedback and text rewards. Quantitative evaluations demonstrate that SCRIPT outperforms prior state-of-the-art methods, with gains across text alignment, motion quality, and physical realism metrics. Furthermore, scaling studies on the 1200-hour MotionMillion dataset demonstrate consistent performance gains with model scaling, highlighting SCRIPT's robust scalability for large-scale pre-training. Our code will be publicly available for future research.

Jiacheng Bao, Haoran Yang, Yucheng Xin et al.

Humanoid robots are expected to execute agile and expressive whole-body motions in real-world settings. Existing text-to-motion generation models are predominantly trained on captured human motion datasets, whose priors assume human biomechanics, actuation, mass distribution, and contact strategies. When such motions are directly retargeted to humanoid robots, the resulting trajectories may satisfy geometric constraints (e.g., joint limits and pose continuity) and appear kinematically reasonable. However, they frequently violate the physical feasibility required for real-world execution. To address these issues, we present PhyGile, a unified framework that closes the loop between robot-native motion generation and General Motion Tracking (GMT). PhyGile performs physics-prefix-guided robot-native motion generation at inference time, directly generating robot-native motions in a 262-dimensional skeletal space with physics-guided prefixes, thereby eliminating inference-time retargeting artifacts and reducing generation-execution discrepancies. Before physics-prefix adaptation, we train the GMT controller with a curriculum-based mixture-of-experts scheme, followed by post-training on unlabeled motion data to improve robustness over large-scale robot motions. During physics-prefix adaptation, the GMT controller is further fine-tuned with generated objectives under physics-derived prefixes, enabling agile and stable execution of complex motions on real robots. Extensive offline and real-robot experiments demonstrate that PhyGile expands the frontier of text-driven humanoid control, enabling stable tracking of agile, highly difficult whole-body motions that go well beyond walking and low-dynamic motions typically achieved by prior methods.

Bin Li, Ruichi Zhang, Han Liang et al.

Humanoid agents are expected to emulate the complex coordination inherent in human social behaviors. However, existing methods are largely confined to single-agent scenarios, overlooking the physically plausible interplay essential for multi-agent interactions. To bridge this gap, we propose InterAgent, the first end-to-end framework for text-driven physics-based multi-agent humanoid control. At its core, we introduce an autoregressive diffusion transformer equipped with multi-stream blocks, which decouples proprioception, exteroception, and action to mitigate cross-modal interference while enabling synergistic coordination. We further propose a novel interaction graph exteroception representation that explicitly captures fine-grained joint-to-joint spatial dependencies to facilitate network learning. Additionally, within it we devise a sparse edge-based attention mechanism that dynamically prunes redundant connections and emphasizes critical inter-agent spatial relations, thereby enhancing the robustness of interaction modeling. Extensive experiments demonstrate that InterAgent consistently outperforms multiple strong baselines, achieving state-of-the-art performance. It enables producing coherent, physically plausible, and semantically faithful multi-agent behaviors from only text prompts. Our code and data will be released to facilitate future research.

Qingcheng Zhao, Pengyu Long, Qixuan Zhang et al.

The synthesis of 3D facial animations from speech has garnered considerable attention. Due to the scarcity of high-quality 4D facial data and well-annotated abundant multi-modality labels, previous methods often suffer from limited realism and a lack of lexible conditioning. We address this challenge through a trilogy. We first introduce Generalized Neural Parametric Facial Asset (GNPFA), an efficient variational auto-encoder mapping facial geometry and images to a highly generalized expression latent space, decoupling expressions and identities. Then, we utilize GNPFA to extract high-quality expressions and accurate head poses from a large array of videos. This presents the M2F-D dataset, a large, diverse, and scan-level co-speech 3D facial animation dataset with well-annotated emotional and style labels. Finally, we propose Media2Face, a diffusion model in GNPFA latent space for co-speech facial animation generation, accepting rich multi-modality guidances from audio, text, and image. Extensive experiments demonstrate that our model not only achieves high fidelity in facial animation synthesis but also broadens the scope of expressiveness and style adaptability in 3D facial animation.

Han Liang, Jiacheng Bao, Ruichi Zhang et al.

We have recently seen tremendous progress in realistic text-to-motion generation. Yet, the existing methods often fail or produce implausible motions with unseen text inputs, which limits the applications. In this paper, we present OMG, a novel framework, which enables compelling motion generation from zero-shot open-vocabulary text prompts. Our key idea is to carefully tailor the pretrain-then-finetune paradigm into the text-to-motion generation. At the pre-training stage, our model improves the generation ability by learning the rich out-of-domain inherent motion traits. To this end, we scale up a large unconditional diffusion model up to 1B parameters, so as to utilize the massive unlabeled motion data up to over 20M motion instances. At the subsequent fine-tuning stage, we introduce motion ControlNet, which incorporates text prompts as conditioning information, through a trainable copy of the pre-trained model and the proposed novel Mixture-of-Controllers (MoC) block. MoC block adaptively recognizes various ranges of the sub-motions with a cross-attention mechanism and processes them separately with the text-token-specific experts. Such a design effectively aligns the CLIP token embeddings of text prompts to various ranges of compact and expressive motion features. Extensive experiments demonstrate that our OMG achieves significant improvements over the state-of-the-art methods on zero-shot text-to-motion generation. Project page: https://tr3e.github.io/omg-page.

Han Liang, Chengyu Huang, Yuecheng Xu et al.

In the realm of Sign Language Translation (SLT), reliance on costly gloss-annotated datasets has posed a significant barrier. Recent advancements in gloss-free SLT methods have shown promise, yet they often largely lag behind gloss-based approaches in terms of translation accuracy. To narrow this performance gap, we introduce LLaVA-SLT, a pioneering Large Multimodal Model (LMM) framework designed to leverage the power of Large Language Models (LLMs) through effectively learned visual language embeddings. Our model is trained through a trilogy. First, we propose linguistic continued pretraining. We scale up the LLM and adapt it to the sign language domain using an extensive corpus dataset, effectively enhancing its textual linguistic knowledge about sign language. Then, we adopt visual contrastive pretraining to align the visual encoder with a large-scale pretrained text encoder. We propose hierarchical visual encoder that learns a robust word-level intermediate representation that is compatible with LLM token embeddings. Finally, we propose visual language tuning. We freeze pretrained models and employ a lightweight trainable MLP connector. It efficiently maps the pretrained visual language embeddings into the LLM token embedding space, enabling downstream SLT task. Our comprehensive experiments demonstrate that LLaVA-SLT outperforms the state-of-the-art methods. By using extra annotation-free data, it even closes to the gloss-based accuracy.

Han Liang, Ziwei Zhan, Weijie Liu et al.

Federated Learning (FL) is a distributed machine learning paradigm that achieves a globally robust model through decentralized computation and periodic model synthesis, primarily focusing on the global model's accuracy over aggregated datasets of all participating clients. Personalized Federated Learning (PFL) instead tailors exclusive models for each client, aiming to enhance the accuracy of clients' individual models on specific local data distributions. Despite of their wide adoption, existing FL and PFL works have yet to comprehensively address the class-imbalance issue, one of the most critical challenges within the realm of data heterogeneity in PFL and FL research. In this paper, we propose FedReMa, an efficient PFL algorithm that can tackle class-imbalance by 1) utilizing an adaptive inter-client co-learning approach to identify and harness different clients' expertise on different data classes throughout various phases of the training process, and 2) employing distinct aggregation methods for clients' feature extractors and classifiers, with the choices informed by the different roles and implications of these model components. Specifically, driven by our experimental findings on inter-client similarity dynamics, we develop critical co-learning period (CCP), wherein we introduce a module named maximum difference segmentation (MDS) to assess and manage task relevance by analyzing the similarities between clients' logits of their classifiers. Outside the CCP, we employ an additional scheme for model aggregation that utilizes historical records of each client's most relevant peers to further enhance the personalization stability. We demonstrate the superiority of our FedReMa in extensive experiments.

Chao Liang, Jianwen Jiang, Wang Liao et al.

Recent advancements in human video generation and animation tasks, driven by diffusion models, have achieved significant progress. However, expressive and realistic human animation remains challenging due to the trade-off between motion naturalness and visual fidelity. To address this, we propose \textbf{AlignHuman}, a framework that combines Preference Optimization as a post-training technique with a divide-and-conquer training strategy to jointly optimize these competing objectives. Our key insight stems from an analysis of the denoising process across timesteps: (1) early denoising timesteps primarily control motion dynamics, while (2) fidelity and human structure can be effectively managed by later timesteps, even if early steps are skipped. Building on this observation, we propose timestep-segment preference optimization (TPO) and introduce two specialized LoRAs as expert alignment modules, each targeting a specific dimension in its corresponding timestep interval. The LoRAs are trained using their respective preference data and activated in the corresponding intervals during inference to enhance motion naturalness and fidelity. Extensive experiments demonstrate that AlignHuman improves strong baselines and reduces NFEs during inference, achieving a 3.3$\times$ speedup (from 100 NFEs to 30 NFEs) with minimal impact on generation quality. Homepage: \href{https://alignhuman.github.io/}{https://alignhuman.github.io/}

Han Liang, Jiahui Zhou, Zicheng Zhou et al.

The escalating adoption of diffusion models for applications such as image generation demands efficient parallel inference techniques to manage their substantial computational cost. However, existing diffusion parallelism inference schemes often underutilize resources in heterogeneous multi-GPU environments, where varying hardware capabilities or background tasks cause workload imbalance. This paper introduces Spatio-Temporal Adaptive Diffusion Inference (STADI), a novel framework to accelerate diffusion model inference in such settings. At its core is a hybrid scheduler that orchestrates fine-grained parallelism across both temporal and spatial dimensions. Temporally, STADI introduces a novel computation-aware step allocator applied after warmup phases, using a least-common-multiple-minimizing quantization technique to reduce denoising steps on slower GPUs and execution synchronization. To further minimize GPU idle periods, STADI executes an elastic patch parallelism mechanism that allocates variably sized image patches to GPUs according to their computational capability, ensuring balanced workload distribution through a complementary spatial mechanism. Extensive experiments on both load-imbalanced and heterogeneous multi-GPU clusters validate STADI's efficacy, demonstrating improved load balancing and mitigation of performance bottlenecks. Compared to patch parallelism, a state-of-the-art diffusion inference framework, our method significantly reduces end-to-end inference latency by up to 45% and significantly improves resource utilization on heterogeneous GPUs.

Jianwen Jiang, Weihong Zeng, Zerong Zheng et al.

Existing video avatar models can produce fluid human animations, yet they struggle to move beyond mere physical likeness to capture a character's authentic essence. Their motions typically synchronize with low-level cues like audio rhythm, lacking a deeper semantic understanding of emotion, intent, or context. To bridge this gap, \textbf{we propose a framework designed to generate character animations that are not only physically plausible but also semantically coherent and expressive.} Our model, \textbf{OmniHuman-1.5}, is built upon two key technical contributions. First, we leverage Multimodal Large Language Models to synthesize a structured textual representation of conditions that provides high-level semantic guidance. This guidance steers our motion generator beyond simplistic rhythmic synchronization, enabling the production of actions that are contextually and emotionally resonant. Second, to ensure the effective fusion of these multimodal inputs and mitigate inter-modality conflicts, we introduce a specialized Multimodal DiT architecture with a novel Pseudo Last Frame design. The synergy of these components allows our model to accurately interpret the joint semantics of audio, images, and text, thereby generating motions that are deeply coherent with the character, scene, and linguistic content. Extensive experiments demonstrate that our model achieves leading performance across a comprehensive set of metrics, including lip-sync accuracy, video quality, motion naturalness and semantic consistency with textual prompts. Furthermore, our approach shows remarkable extensibility to complex scenarios, such as those involving multi-person and non-human subjects. Homepage: \href{https://omnihuman-lab.github.io/v1_5/}

Team Seedance, Heyi Chen, Siyan Chen et al.

Recent strides in video generation have paved the way for unified audio-visual generation. In this work, we present Seedance 1.5 pro, a foundational model engineered specifically for native, joint audio-video generation. Leveraging a dual-branch Diffusion Transformer architecture, the model integrates a cross-modal joint module with a specialized multi-stage data pipeline, achieving exceptional audio-visual synchronization and superior generation quality. To ensure practical utility, we implement meticulous post-training optimizations, including Supervised Fine-Tuning (SFT) on high-quality datasets and Reinforcement Learning from Human Feedback (RLHF) with multi-dimensional reward models. Furthermore, we introduce an acceleration framework that boosts inference speed by over 10X. Seedance 1.5 pro distinguishes itself through precise multilingual and dialect lip-syncing, dynamic cinematic camera control, and enhanced narrative coherence, positioning it as a robust engine for professional-grade content creation. Seedance 1.5 pro is now accessible on Volcano Engine at https://console.volcengine.com/ark/region:ark+cn-beijing/experience/vision?type=GenVideo.

Yannan He, Anqi Pang, Xin Chen et al.

Capturing challenging human motions is critical for numerous applications, but it suffers from complex motion patterns and severe self-occlusion under the monocular setting. In this paper, we propose ChallenCap -- a template-based approach to capture challenging 3D human motions using a single RGB camera in a novel learning-and-optimization framework, with the aid of multi-modal references. We propose a hybrid motion inference stage with a generation network, which utilizes a temporal encoder-decoder to extract the motion details from the pair-wise sparse-view reference, as well as a motion discriminator to utilize the unpaired marker-based references to extract specific challenging motion characteristics in a data-driven manner. We further adopt a robust motion optimization stage to increase the tracking accuracy, by jointly utilizing the learned motion details from the supervised multi-modal references as well as the reliable motion hints from the input image reference. Extensive experiments on our new challenging motion dataset demonstrate the effectiveness and robustness of our approach to capture challenging human motions.