Iterative Structured Pruning for Large Language Models with Multi-Domain Calibration

Large Language Models (LLMs) have achieved remarkable success across a wide spectrum of natural language processing tasks. However, their ever-growing scale introduces significant barriers to real-world deployment, including substantial computational overhead, memory footprint, and inference latency. While model pruning presents a viable solution to these challenges, existing unstructured pruning techniques often yield irregular sparsity patterns that necessitate specialized hardware or software support. In this work, we explore structured pruning, which eliminates entire architectural components and maintains compatibility with standard hardware accelerators. We introduce a novel structured pruning framework that leverages a hybrid multi-domain calibration set and an iterative calibration strategy to effectively identify and remove redundant channels. Extensive experiments on various models across diverse downstream tasks show that our approach achieves significant compression with minimal performance degradation.

Quantized SO(3)-Equivariant Graph Neural Networks for Efficient Molecular Property Prediction

Deploying 3D graph neural networks (GNNs) that are equivariant to 3D rotations (the group SO(3)) on edge devices is challenging due to their high computational cost. This paper addresses the problem by compressing and accelerating an SO(3)-equivariant GNN using low-bit quantization techniques. Specifically, we introduce three innovations for quantized equivariant transformers: (1) a magnitude-direction decoupled quantization scheme that separately quantizes the norm and orientation of equivariant (vector) features, (2) a branch-separated quantization-aware training strategy that treats invariant and equivariant feature channels differently in an attention-based $SO(3)$-GNN, and (3) a robustness-enhancing attention normalization mechanism that stabilizes low-precision attention computations. Experiments on the QM9 and rMD17 molecular benchmarks demonstrate that our 8-bit models achieve accuracy on energy and force predictions comparable to full-precision baselines with markedly improved efficiency. We also conduct ablation studies to quantify the contribution of each component to maintain accuracy and equivariance under quantization, using the Local error of equivariance (LEE) metric. The proposed techniques enable the deployment of symmetry-aware GNNs in practical chemistry applications with 2.37--2.73x faster inference and 4x smaller model size, without sacrificing accuracy or physical symmetry.

TrimTokenator-LC: Towards Adaptive Visual Token Pruning for Large Multimodal Models with Long Contexts

Large Multimodal Models (LMMs) have proven effective on various tasks. They typically encode visual inputs into Original Model sequences of tokens, which are then concatenated with textual tokens and jointly processed by the language model. However, the growing number of visual tokens greatly increases inference cost. Visual token pruning has emerged as a promising solution. However, existing methods often overlook scenarios involving long context inputs with multiple images. In this paper, we analyze the challenges of visual token pruning in long context, multi-image settings and introduce an adaptive pruning method tailored for such scenarios. We decompose redundancy into intra-image and inter-image components and quantify them through intra-image diversity and inter-image variation, which jointly guide dynamic budget allocation. Our approach consists of two stages. The intra-image stage allocates each image a content-aware token budget and greedily selects its most representative tokens. The inter-image stage performs global diversity filtering to form a candidate pool and then applies a Pareto selection procedure that balances diversity with text alignment. Extensive experiments show that our approach can reduce up to 80% of visual tokens while maintaining performance in long context settings.

HY-Motion 1.0: Scaling Flow Matching Models for Text-To-Motion Generation

We present HY-Motion 1.0, a series of state-of-the-art, large-scale, motion generation models capable of generating 3D human motions from textual descriptions. HY-Motion 1.0 represents the first successful attempt to scale up Diffusion Transformer (DiT)-based flow matching models to the billion-parameter scale within the motion generation domain, delivering instruction-following capabilities that significantly outperform current open-source benchmarks. Uniquely, we introduce a comprehensive, full-stage training paradigm -- including large-scale pretraining on over 3,000 hours of motion data, high-quality fine-tuning on 400 hours of curated data, and reinforcement learning from both human feedback and reward models -- to ensure precise alignment with the text instruction and high motion quality. This framework is supported by our meticulous data processing pipeline, which performs rigorous motion cleaning and captioning. Consequently, our model achieves the most extensive coverage, spanning over 200 motion categories across 6 major classes. We release HY-Motion 1.0 to the open-source community to foster future research and accelerate the transition of 3D human motion generation models towards commercial maturity.

ByteLoom: Weaving Geometry-Consistent Human-Object Interactions through Progressive Curriculum Learning

Human-object interaction (HOI) video generation has garnered increasing attention due to its promising applications in digital humans, e-commerce, advertising, and robotics imitation learning. However, existing methods face two critical limitations: (1) a lack of effective mechanisms to inject multi-view information of the object into the model, leading to poor cross-view consistency, and (2) heavy reliance on fine-grained hand mesh annotations for modeling interaction occlusions. To address these challenges, we introduce ByteLoom, a Diffusion Transformer (DiT)-based framework that generates realistic HOI videos with geometrically consistent object illustration, using simplified human conditioning and 3D object inputs. We first propose an RCM-cache mechanism that leverages Relative Coordinate Maps (RCM) as a universal representation to maintain object's geometry consistency and precisely control 6-DoF object transformations in the meantime. To compensate HOI dataset scarcity and leverage existing datasets, we further design a training curriculum that enhances model capabilities in a progressive style and relaxes the demand of hand mesh. Extensive experiments demonstrate that our method faithfully preserves human identity and the object's multi-view geometry, while maintaining smooth motion and object manipulation.

PTalker: Personalized Speech-Driven 3D Talking Head Animation via Style Disentanglement and Modality Alignment

Speech-driven 3D talking head generation aims to produce lifelike facial animations precisely synchronized with speech. While considerable progress has been made in achieving high lip-synchronization accuracy, existing methods largely overlook the intricate nuances of individual speaking styles, which limits personalization and realism. In this work, we present a novel framework for personalized 3D talking head animation, namely "PTalker". This framework preserves speaking style through style disentanglement from audio and facial motion sequences and enhances lip-synchronization accuracy through a three-level alignment mechanism between audio and mesh modalities. Specifically, to effectively disentangle style and content, we design disentanglement constraints that encode driven audio and motion sequences into distinct style and content spaces to enhance speaking style representation. To improve lip-synchronization accuracy, we adopt a modality alignment mechanism incorporating three aspects: spatial alignment using Graph Attention Networks to capture vertex connectivity in the 3D mesh structure, temporal alignment using cross-attention to capture and synchronize temporal dependencies, and feature alignment by top-k bidirectional contrastive losses and KL divergence constraints to ensure consistency between speech and mesh modalities. Extensive qualitative and quantitative experiments on public datasets demonstrate that PTalker effectively generates realistic, stylized 3D talking heads that accurately match identity-specific speaking styles, outperforming state-of-the-art methods. The source code and supplementary videos are available at: PTalker.

Self-motion as a structural prior for coherent and robust formation of cognitive maps

Most computational accounts of cognitive maps assume that stability is achieved primarily through sensory anchoring, with self-motion contributing to incremental positional updates only. However, biological spatial representations often remain coherent even when sensory cues degrade or conflict, suggesting that self-motion may play a deeper organizational role. Here, we show that self-motion can act as a structural prior that actively organizes the geometry of learned cognitive maps. We embed a path-integration-based motion prior in a predictive-coding framework, implemented using a capacity-efficient, brain-inspired recurrent mechanism combining spiking dynamics, analog modulation and adaptive thresholds. Across highly aliased, dynamically changing and naturalistic environments, this structural prior consistently stabilizes map formation, improving local topological fidelity, global positional accuracy and next-step prediction under sensory ambiguity. Mechanistic analyses reveal that the motion prior itself encodes geometrically precise trajectories under tight constraints of internal states and generalizes zero-shot to unseen environments, outperforming simpler motion-based constraints. Finally, deployment on a quadrupedal robot demonstrates that motion-derived structural priors enhance online landmark-based navigation under real-world sensory variability. Together, these results reframe self-motion as an organizing scaffold for coherent spatial representations, showing how brain-inspired principles can systematically strengthen spatial intelligence in embodied artificial agents.

CETCAM: Camera-Controllable Video Generation via Consistent and Extensible Tokenization

Achieving precise camera control in video generation remains challenging, as existing methods often rely on camera pose annotations that are difficult to scale to large and dynamic datasets and are frequently inconsistent with depth estimation, leading to train-test discrepancies. We introduce CETCAM, a camera-controllable video generation framework that eliminates the need for camera annotations through a consistent and extensible tokenization scheme. CETCAM leverages recent advances in geometry foundation models, such as VGGT, to estimate depth and camera parameters and converts them into unified, geometry-aware tokens. These tokens are seamlessly integrated into a pretrained video diffusion backbone via lightweight context blocks. Trained in two progressive stages, CETCAM first learns robust camera controllability from diverse raw video data and then refines fine-grained visual quality using curated high-fidelity datasets. Extensive experiments across multiple benchmarks demonstrate state-of-the-art geometric consistency, temporal stability, and visual realism. Moreover, CETCAM exhibits strong adaptability to additional control modalities, including inpainting and layout control, highlighting its flexibility beyond camera control. The project page is available at https://sjtuytc.github.io/CETCam_project_page.github.io/.

Fast and Accurate Causal Parallel Decoding using Jacobi Forcing

Multi-token generation has emerged as a promising paradigm for accelerating transformer-based large model inference. Recent efforts primarily explore diffusion Large Language Models (dLLMs) for parallel decoding to reduce inference latency. To achieve AR-level generation quality, many techniques adapt AR models into dLLMs to enable parallel decoding. However, they suffer from limited speedup compared to AR models due to a pretrain-to-posttrain mismatch. Specifically, the masked data distribution in post-training deviates significantly from the real-world data distribution seen during pretraining, and dLLMs rely on bidirectional attention, which conflicts with the causal prior learned during pretraining and hinders the integration of exact KV cache reuse. To address this, we introduce Jacobi Forcing, a progressive distillation paradigm where models are trained on their own generated parallel decoding trajectories, smoothly shifting AR models into efficient parallel decoders while preserving their pretrained causal inference property. The models trained under this paradigm, Jacobi Forcing Model, achieves 3.8x wall-clock speedup on coding and math benchmarks with minimal loss in performance. Based on Jacobi Forcing Models' trajectory characteristics, we introduce multi-block decoding with rejection recycling, which enables up to 4.5x higher token acceptance count per iteration and nearly 4.0x wall-clock speedup, effectively trading additional compute for lower inference latency. Our code is available at https://github.com/hao-ai-lab/JacobiForcing.

Efficient-DLM: From Autoregressive to Diffusion Language Models, and Beyond in Speed

Diffusion language models (dLMs) have emerged as a promising paradigm that enables parallel, non-autoregressive generation, but their learning efficiency lags behind that of autoregressive (AR) language models when trained from scratch. To this end, we study AR-to-dLM conversion to transform pretrained AR models into efficient dLMs that excel in speed while preserving AR models' task accuracy. We achieve this by identifying limitations in the attention patterns and objectives of existing AR-to-dLM methods and then proposing principles and methodologies for more effective AR-to-dLM conversion. Specifically, we first systematically compare different attention patterns and find that maintaining pretrained AR weight distributions is critical for effective AR-to-dLM conversion. As such, we introduce a continuous pretraining scheme with a block-wise attention pattern, which remains causal across blocks while enabling bidirectional modeling within each block. We find that this approach can better preserve pretrained AR models' weight distributions than fully bidirectional modeling, in addition to its known benefit of enabling KV caching, and leads to a win-win in accuracy and efficiency. Second, to mitigate the training-test gap in mask token distributions (uniform vs. highly left-to-right), we propose a position-dependent token masking strategy that assigns higher masking probabilities to later tokens during training to better mimic test-time behavior. Leveraging this framework, we conduct extensive studies of dLMs' attention patterns, training dynamics, and other design choices, providing actionable insights into scalable AR-to-dLM conversion. These studies lead to the Efficient-DLM family, which outperforms state-of-the-art AR models and dLMs, e.g., our Efficient-DLM 8B achieves +5.4%/+2.7% higher accuracy with 4.5x/2.7x higher throughput compared to Dream 7B and Qwen3 4B, respectively.