AtlasVid: Efficient Ultra-High-Resolution Long Video Generation via Decoupled Global-Local Modeling
- URL: http://arxiv.org/abs/2605.16649v1
- Date: Fri, 15 May 2026 21:39:46 GMT
- Title: AtlasVid: Efficient Ultra-High-Resolution Long Video Generation via Decoupled Global-Local Modeling
- Abstract summary: We propose AtlaVid, a decoupled global-local framework for efficient ultra-high-resolution long video generation.<n>AtlaVid generates a low-resolution and low-FPS global semantic proxy via temporally scaled RoPE, thereby extending the temporal horizon without increasing the training token count.<n>Experiments show that AtlaVid substantially improves the efficiency of ultra-high-resolution long video generation, achieving high-quality UHR long video generation with 6x>0.9 speed and less training cost and even better performance than native 4K video generators.
- Score: 44.836661735003474
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: Recent diffusion-based video generators have achieved remarkable visual fidelity and prompt controllability, yet scaling them to ultra-high-resolution (UHR) long videos remains prohibitively expensive. The difficulty is especially pronounced for long single-shot generation where a continuous scene must preserve global temporal coherence, and fine-grained spatial details without relying on clip transitions or autoregressive shot stitching. In this work, we revisit this challenge from the perspective of decoupled modeling. We argue that existing video diffusion models already encode strong local visual priors, while the main bottleneck lies in efficiently extending global spatiotemporal modeling as resolution and duration increase. Based on this insight, we propose AtlaVid, a decoupled global-local framework for efficient UHR long video generation. AtlaVid first generates a low-resolution and low-FPS global semantic proxy via temporally scaled RoPE, thereby extending the temporal horizon without increasing the training token count. Guided by this proxy, a high-resolution detail branch performs joint denoising with hierarchical locality-preserving attention. Reordered spatiotemporal windows preserve geometric locality and asymmetric global-local attention injects aligned semantic guidance and preserves the model's pretrained ability. This design enables resolution-agnostic training: the model is trained only at 720P with lightweight LoRA adaptation, yet generalizes directly to 4K and beyond for longer (>10s) video synthesis. Experiments show that AtlaVid substantially improves the efficiency of ultra-high-resolution long video generation, achieving high-quality UHR long video generation with 60.9x speed up and significantly less training cost and even better performance than native 4K video generators.
Related papers
- DCARL: A Divide-and-Conquer Framework for Autoregressive Long-Trajectory Video Generation [10.575655795360937]
DCARL is a novel divide-and-conquer, autoregressive framework that combines the structural stability of the divide-and-conquer scheme with the high-fidelity generation of VDMs.<n>Our method achieves superior performance in both visual quality and camera adherence compared to state-of-the-art autoregressive and divide-and-conquer baselines.
arXiv Detail & Related papers (2026-03-25T22:07:09Z) - FrescoDiffusion: 4K Image-to-Video with Prior-Regularized Tiled Diffusion [46.49480145234397]
We introduce FrescoDiffusion, a training-free method for coherent large-format I2V generation from a single image.<n>For 4K generation, we compute per-tile noise predictions and fuse them with this reference at every diffusion timestep.<n>Experiments on the VBench-I2V dataset and our proposed fresco I2V dataset show improved global consistency and fidelity over tiled baselines.
arXiv Detail & Related papers (2026-03-18T10:02:37Z) - Towards Holistic Modeling for Video Frame Interpolation with Auto-regressive Diffusion Transformers [95.68243351895107]
We propose a holistic, video-centric paradigm named textbfLocal textbfDiffusion textbfForcing for textbfVideo textbfFrame textbfInterpolation (LDF-VFI)<n>Our framework is built upon an auto-regressive diffusion transformer that models the entire video sequence to ensure long-range temporal coherence.<n>LDF-VFI achieves state-of-the-art performance on challenging long-sequence benchmarks, demonstrating superior per
arXiv Detail & Related papers (2026-01-21T12:58:52Z) - UltraGen: High-Resolution Video Generation with Hierarchical Attention [62.99161115650818]
UltraGen is a novel video generation framework that enables i) efficient and ii) end-to-end native high-resolution video synthesis.<n>We show that UltraGen can effectively scale pre-trained low-resolution video models to 1080P and even 4K resolution for the first time.
arXiv Detail & Related papers (2025-10-21T16:23:21Z) - Towards Robust and Generalizable Continuous Space-Time Video Super-Resolution with Events [71.2439653098351]
Continuous space-time video super-STVSR has garnered increasing interest for its capability to reconstruct high-resolution and high-frame-rate videos at arbitrary temporal scales.<n>We present EvEnhancer, a novel approach that marries unique properties of high temporal and high dynamic range encapsulated in event streams.<n>Our method achieves state-of-the-art performance on both synthetic and real-world datasets, while maintaining generalizability at OOD scales.
arXiv Detail & Related papers (2025-10-04T15:23:07Z) - Tuning-Free Long Video Generation via Global-Local Collaborative Diffusion [22.988212617368095]
We propose GLC-Diffusion, a tuning-free method for long video generation.<n>It models the long video denoising process by establishing Global-Local Collaborative Denoising.<n>We also propose a Video Motion Consistency Refinement (VMCR) module that computes the gradient of pixel-wise and frequency-wise losses.
arXiv Detail & Related papers (2025-01-08T05:49:39Z) - FreeLong: Training-Free Long Video Generation with SpectralBlend Temporal Attention [57.651429116402554]
This paper investigates a straightforward and training-free approach to extend an existing short video diffusion model for consistent long video generation.
We find that directly applying the short video diffusion model to generate long videos can lead to severe video quality degradation.
Motivated by this, we propose a novel solution named FreeLong to balance the frequency distribution of long video features during the denoising process.
arXiv Detail & Related papers (2024-07-29T11:52:07Z) - Upscale-A-Video: Temporal-Consistent Diffusion Model for Real-World
Video Super-Resolution [65.91317390645163]
Upscale-A-Video is a text-guided latent diffusion framework for video upscaling.
It ensures temporal coherence through two key mechanisms: locally, it integrates temporal layers into U-Net and VAE-Decoder, maintaining consistency within short sequences.
It also offers greater flexibility by allowing text prompts to guide texture creation and adjustable noise levels to balance restoration and generation.
arXiv Detail & Related papers (2023-12-11T18:54:52Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.