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
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use continuous VAE tokenizer and sampler instead of VQ-VAE (ablation) | next-set AR model · ImageNet 256x256 class-conditional generation | FID: 2.49 -> 2.05 | Improved |
| use continuous VAE tokenizer instead of VQ-VAE (ablation) | next-set AR model · UCF-101 zero-shot text-to-video | FVD: 314 -> 247 | Improved |
| use continuous VAE tokenizer instead of VQ-VAE (ablation) | next-set AR model · VBench | temporal flickering: 97.87 -> 98.92 | Improved |
| use continuous VAE tokenizer instead of VQ-VAE (ablation) | next-set AR model · VBench | motion smoothness: 97.85 -> 98.67 | Improved |
| increase number of sets K from 8 to 16 (ablation) | next-set AR model · ImageNet 256x256 class-conditional generation | FID: 4.46 -> 2.73 | Improved |
| increase number of sets K from 16 to 32 (ablation) | next-set AR model · ImageNet 256x256 class-conditional generation | FID: 2.73 -> 2.05 | Improved |
| increase number of sets K from 32 to 64 (ablation) | next-set AR model · ImageNet 256x256 class-conditional generation | FID: 2.05 -> 1.86 | Improved |
| increase number of sets K from 32 to 64 (ablation) | next-set AR model · ImageNet 256x256 class-conditional generation | latency (s): 13.11 -> 20.08 | Hurt |
| reduce number of sets K from 32 to 8 (ablation) | next-set AR model · ImageNet 256x256 class-conditional generation | latency (s): 13.11 -> 6.45 | Improved |
| scale next-set AR model from 343M to 775M parameters | next-set AR model · UCF-101 zero-shot text-to-video | FVD: 247 -> 204 | Improved |
| scale next-set AR model from 343M to 775M parameters | next-set AR model · MSR-VTT zero-shot text-to-video | CLIPSIM: 0.2942 -> 0.3092 | Improved |
| scale next-set AR model from 343M to 775M parameters | next-set AR model · UCF-101 zero-shot text-to-video | latency (s/frame): 1.1 -> 1.6 | Hurt |
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.
- VideoPoet — VideoPoet: A Large Language Model for Zero-Shot Video Generation (ICML 2024) · better on 5 of 5 shared results
- CogVideoX-2B — CogVideoX: Text-to-Video Diffusion Models with An Expert Transformer (ICLR 2025) · better on 0 of 1 shared result
- Show-1 — Show-1: Marrying Pixel and Latent Diffusion Models for Text-to-Video Generation (IJCV 2024) · better on 1 of 1 shared result
- Make-A-Video — Make-A-Video: Text-to-Video Generation without Text-Video Data (ICLR 2023) · better on 1 of 1 shared result
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
- From Slow Bidirectional to Fast Autoregressive Video Diffusion Models (CVPR 2025 · similarity 0.89)
- Training-Free Efficient Video Generation via Dynamic Token Carving (NeurIPS 2025 · similarity 0.89)
- OSV: One Step is Enough for High-Quality Image to Video Generation (CVPR 2025 · similarity 0.89)
- Video Prediction by Modeling Videos as Continuous Multi-Dimensional Processes (CVPR 2024 · similarity 0.89)
- AR-Diffusion: Asynchronous Video Generation with Auto-Regressive Diffusion (CVPR 2025 · similarity 0.89)
- FreeLong: Training-Free Long Video Generation with SpectralBlend Temporal Attention (NeurIPS 2024 · similarity 0.88)
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.
