LeituraPesquisaRadarFramework de investimento
Entrar / Cadastrar
← Voltar para a leitura
NotíciasTópico · Optica因果视频模型Material1 min de leituraEditorial DisconfirmAI

Preprint: Conditional Residual Prediction Helps Causal Video Model Reach 82.78 on VBench Without Teacher Distillation

New training method enables a 2B-parameter causal video model to score 82.78 on VBench, approaching bidirectional performance.

2026-10-09 12:00 GMT+8

Key Finding: A new training method called Conditional Residual Prediction (CRP) enabled the 2-billion-parameter causal video diffusion model Optica to achieve a score of 82.78 on the VBench benchmark, using only approximately 15 million training videos.

Context: Traditional causal video models, suitable for streaming and long-form generation due to their autoregressive nature, typically yield lower quality than bidirectional models of the same size. Existing solutions often rely on knowledge distillation from large bidirectional teacher models, which is complex and computationally expensive.

Result: CRP reduces the model's over-reliance on history by predicting the target first and adding historical conditions as residuals. Experiments show this approach nearly closes the 6.14-point gap with bidirectional models trained under the same setup, without requiring any bidirectional video model during training.

Limitations: Results are self-reported in a preprint and have not yet been independently reproduced. The validation focuses on 480p, 5-second video generation tasks.

Fontes

  • arxiv.org↗
Tópico

Optica因果视频模型

【背景】因果视频扩散模型虽适合流媒体生成,但传统上需依赖双向教师模型蒸馏以弥补质量差距。2026年10月8日,Bowen Zheng等人在arXiv发布预印本,提出“条件残差预测”(CRP)训练方法,并推出20亿参数的因果视频模型Optic…

Entrar / Cadastrar
LeituraPesquisaRadarFramework de investimento

2 etapas1 peçasAtualizado 2026-10-10 16:22

Como chegamos aqui

  1. 2026-10-08论文《Conditional Residual Prediction: Improving Autoregressive Video Diffusion without a Bidirectional Teacher》提交至arXiv,作者包括Bowen Zheng、Zhiguang Liu等。arxiv.org ↗
  2. 2026-10-08研究团队提出CRP方法,训练出2B参数因果视频模型Optica,实现5秒480p视频自回归生成,VBench得分82.78。arxiv.org ↗

O que ainda está indefinido

  • Optica模型是否已开源代码或权重?
  • CRP方法在更长时长(如超过5秒)的视频生成中的稳定性如何?
  • 该方法与现有的其他免教师蒸馏方案(如Self-Resampling)相比,计算效率具体提升多少?
Ler o tópico completo →
Continuar pesquisando →