Fast Autoregressive Video Generation with Diagonal Decoding

• URL: http://arxiv.org/abs/2503.14070v1
• Date: Tue, 18 Mar 2025 09:42:55 GMT
• Title: Fast Autoregressive Video Generation with Diagonal Decoding
• Authors: Yang Ye, Junliang Guo, Haoyu Wu, Tianyu He, Tim Pearce, Tabish Rashid, Katja Hofmann, Jiang Bian,
• Abstract summary: Diagonal Decoding (DiagD) is a training-free inference acceleration algorithm for autoregressively pre-trained models.<n>Our method generates tokens along diagonal paths in the spatial-temporal token grid, enabling parallel decoding within each frame.<n>DiagD achieves up to $10times$ speedup compared to naive sequential decoding, while maintaining comparable visual fidelity.
• Score: 34.90521536645348
• License: http://creativecommons.org/licenses/by-nc-nd/4.0/
• Abstract: Autoregressive Transformer models have demonstrated impressive performance in video generation, but their sequential token-by-token decoding process poses a major bottleneck, particularly for long videos represented by tens of thousands of tokens. In this paper, we propose Diagonal Decoding (DiagD), a training-free inference acceleration algorithm for autoregressively pre-trained models that exploits spatial and temporal correlations in videos. Our method generates tokens along diagonal paths in the spatial-temporal token grid, enabling parallel decoding within each frame as well as partially overlapping across consecutive frames. The proposed algorithm is versatile and adaptive to various generative models and tasks, while providing flexible control over the trade-off between inference speed and visual quality. Furthermore, we propose a cost-effective finetuning strategy that aligns the attention patterns of the model with our decoding order, further mitigating the training-inference gap on small-scale models. Experiments on multiple autoregressive video generation models and datasets demonstrate that DiagD achieves up to $10\times$ speedup compared to naive sequential decoding, while maintaining comparable visual fidelity.

Related papers

• Generating, Fast and Slow: Scalable Parallel Video Generation with Video Interface Networks [21.710127132217526]
  We introduce a new paradigm called Video Interface Networks (VINs), which augment DiTs with an abstraction module to enable parallel inference of video chunks. VINs encode global semantics from the noisy input of local chunks and the encoded representations, in turn, guide DiTs in denoising chunks in parallel. Our approach attains state-of-the-art motion smoothness while using 25-40% fewer FLOPs than full generation.
  arXiv  Detail & Related papers  (2025-03-21T21:13:02Z)
• Token-Efficient Long Video Understanding for Multimodal LLMs [101.70681093383365]
  STORM is a novel architecture incorporating a dedicated temporal encoder between the image encoder and the Video-LLMs.<n>We show that STORM achieves state-of-the-art results across various long video understanding benchmarks.
  arXiv  Detail & Related papers  (2025-03-06T06:17:38Z)
• Raccoon: Multi-stage Diffusion Training with Coarse-to-Fine Curating Videos [15.781862060265519]
  CFC-VIDS-1M is a high-quality video dataset constructed through a systematic coarse-to-fine curation pipeline.<n>We develop RACCOON, a transformer-based architecture with decoupled spatial-temporal attention mechanisms.
  arXiv  Detail & Related papers  (2025-02-28T18:56:35Z)
• Optical-Flow Guided Prompt Optimization for Coherent Video Generation [51.430833518070145]
  We propose a framework called MotionPrompt that guides the video generation process via optical flow. We optimize learnable token embeddings during reverse sampling steps by using gradients from a trained discriminator applied to random frame pairs. This approach allows our method to generate visually coherent video sequences that closely reflect natural motion dynamics, without compromising the fidelity of the generated content.
  arXiv  Detail & Related papers  (2024-11-23T12:26:52Z)
• Accelerating Auto-regressive Text-to-Image Generation with Training-free Speculative Jacobi Decoding [60.188309982690335]
  We propose a training-free probabilistic parallel decoding algorithm, Speculative Jacobi Decoding (SJD)<n>SJD accelerates the inference of auto-regressive text-to-image generation while maintaining the randomness in sampling-based token decoding.<n>Specifically, SJD facilitates the model to predict multiple tokens at each step and accepts tokens based on the probabilistic criterion.
  arXiv  Detail & Related papers  (2024-10-02T16:05:27Z)
• RAVEN: Rethinking Adversarial Video Generation with Efficient Tri-plane Networks [93.18404922542702]
  We present a novel video generative model designed to address long-term spatial and temporal dependencies. Our approach incorporates a hybrid explicit-implicit tri-plane representation inspired by 3D-aware generative frameworks. Our model synthesizes high-fidelity video clips at a resolution of $256times256$ pixels, with durations extending to more than $5$ seconds at a frame rate of 30 fps.
  arXiv  Detail & Related papers  (2024-01-11T16:48:44Z)
• Diffusion Glancing Transformer for Parallel Sequence to Sequence Learning [52.72369034247396]
  We propose the diffusion glancing transformer, which employs a modality diffusion process and residual glancing sampling. DIFFGLAT achieves better generation accuracy while maintaining fast decoding speed compared with both autoregressive and non-autoregressive models.
  arXiv  Detail & Related papers  (2022-12-20T13:36:25Z)
• Autoencoding Video Latents for Adversarial Video Generation [0.0]
  AVLAE is a two stream latent autoencoder where the video distribution is learned by adversarial training. We demonstrate that our approach learns to disentangle motion and appearance codes even without the explicit structural composition in the generator.
  arXiv  Detail & Related papers  (2022-01-18T11:42:14Z)
• ViViT: A Video Vision Transformer [75.74690759089529]
  We present pure-transformer based models for video classification. Our model extracts-temporal tokens from the input video, which are then encoded by a series of transformer layers. We show how we can effectively regularise the model during training and leverage pretrained image models to be able to train on comparatively small datasets.
  arXiv  Detail & Related papers  (2021-03-29T15:27:17Z)
• PGT: A Progressive Method for Training Models on Long Videos [45.935259079953255]
  Main-stream method is to split a raw video into clips, leading to incomplete temporal information flow. Inspired by natural language processing techniques dealing with long sentences, we propose to treat videos as serial fragments satisfying Markov property. We empirically demonstrate that it yields significant performance improvements on different models and datasets.
  arXiv  Detail & Related papers  (2021-03-21T06:15:20Z)
• An Emerging Coding Paradigm VCM: A Scalable Coding Approach Beyond Feature and Signal [99.49099501559652]
  Video Coding for Machine (VCM) aims to bridge the gap between visual feature compression and classical video coding. We employ a conditional deep generation network to reconstruct video frames with the guidance of learned motion pattern. By learning to extract sparse motion pattern via a predictive model, the network elegantly leverages the feature representation to generate the appearance of to-be-coded frames.
  arXiv  Detail & Related papers  (2020-01-09T14:18:18Z)

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.