AI Paper Atlas

Research map / Video / Video generation

Unified Video Generation via Next-Set Prediction in Continuous Domain

Zhanzhou Feng, Qingpei Guo, Xinyu Xiao, Ruihan Xu, Ming Yang, Shiliang Zhang

ICCV 2025 · Video › Video generation · approach: Efficient video diffusion models

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use continuous VAE tokenizer and sampler instead of VQ-VAE (ablation)next-set AR model · ImageNet 256x256 class-conditional generationFID: 2.49 -> 2.05Improved
use continuous VAE tokenizer instead of VQ-VAE (ablation)next-set AR model · UCF-101 zero-shot text-to-videoFVD: 314 -> 247Improved
use continuous VAE tokenizer instead of VQ-VAE (ablation)next-set AR model · VBenchtemporal flickering: 97.87 -> 98.92Improved
use continuous VAE tokenizer instead of VQ-VAE (ablation)next-set AR model · VBenchmotion smoothness: 97.85 -> 98.67Improved
increase number of sets K from 8 to 16 (ablation)next-set AR model · ImageNet 256x256 class-conditional generationFID: 4.46 -> 2.73Improved
increase number of sets K from 16 to 32 (ablation)next-set AR model · ImageNet 256x256 class-conditional generationFID: 2.73 -> 2.05Improved
increase number of sets K from 32 to 64 (ablation)next-set AR model · ImageNet 256x256 class-conditional generationFID: 2.05 -> 1.86Improved
increase number of sets K from 32 to 64 (ablation)next-set AR model · ImageNet 256x256 class-conditional generationlatency (s): 13.11 -> 20.08Hurt
reduce number of sets K from 32 to 8 (ablation)next-set AR model · ImageNet 256x256 class-conditional generationlatency (s): 13.11 -> 6.45Improved
scale next-set AR model from 343M to 775M parametersnext-set AR model · UCF-101 zero-shot text-to-videoFVD: 247 -> 204Improved
scale next-set AR model from 343M to 775M parametersnext-set AR model · MSR-VTT zero-shot text-to-videoCLIPSIM: 0.2942 -> 0.3092Improved
scale next-set AR model from 343M to 775M parametersnext-set AR model · UCF-101 zero-shot text-to-videolatency (s/frame): 1.1 -> 1.6Hurt

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 4 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy