AI Paper Atlas

Research map / Video / Video generation

Pusa V1.0: Unlocking Temporal Control in Pretrained Video Diffusion Models via Vectorized Timestep Adaptation

Yaofang Liu, Yumeng REN, Aitor Artola, Yuxuan Hu, Xiaodong Cun, Xiaotong Zhao, Alan Zhao, Raymond H. Chan, Suiyun Zhang, Rui Liu, Dandan Tu, Jean-michel Morel

ICLR 2026 · Video › Video generation · approach: Efficient video diffusion models · 5 citations (OpenAlex)

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
replace Wan-I2V fine-tuning with vectorized timestep adaptation (Pusa)Wan2.1-T2V-14B · VBench-I2VTotal score (%): 86.86 -> 87.32Improved
replace Wan-I2V fine-tuning with PusaWan2.1-T2V-14B · VBench-I2VQuality score (%): 80.82 -> 79.80Hurt
replace Wan-I2V training with VTA at equal 900-step budget (480p) (ablation)Wan2.1-T2V-14B · VBench-I2V subsetTotal score (%): 73.21 -> 87.09Improved
use fully random per-frame timesteps instead of I2V-style fixed first frame (ablation)Pusa · VBench-I2V subsetTotal score (%): 73.27 -> 87.69Improved
use fully random timesteps instead of PTSS p=0.2 (ablation)Pusa · VBench-I2V subsetTotal score (%): 84.74 -> 87.69Improved
use fully random timesteps instead of PTSS p=0.8 (ablation)Pusa · VBench-I2V subsetTotal score (%): 86.49 -> 87.69Improved
increase LoRA rank from 256 to 512 (ablation)Pusa · VBench-I2V subsetTotal score (%): 83.12 -> 86.46Improved
increase inference steps from 10 to 20 (ablation)Pusa · VBench-I2V subsetTotal score (%): 87.69 -> 87.84No clear effect
increase inference steps from 2 to 10 (ablation)Pusa · VBench-I2V subsetTotal score (%): 79.92 -> 87.69Improved
train beyond 900 iterations (ablation)Pusa LoRA · VBench-I2V subsetTotal score (%): 87.69 -> 87.32No clear effect
swap base model Wan2.1 for Wan2.2 (ablation)Pusa · VBench-I2V subsetTotal score (%): 87.69 -> 87.69No clear effect
use full fine-tuning instead of LoRA (ablation)Pusa · VBench-I2V subsetTotal score (%): 84.06 vs 87.09No clear effect

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