AI Paper Atlas

Research map / Video / Video generation

Temporal-aware Flow Matching for Video Generation with Temporally Coherent Motion

Zirui Pan, Xin Wang, Yipeng Zhang, Yuwei Zhou, Wenwu Zhu

ICML 2026 · Video › Video generation · approach: Autoregressive long video generation

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
fine-tune with TFM flow instead of standard FM flow (ablation)Wan2.1-T2V-14B · VideoJam-BenchVBench motion score: 0.815 -> 0.961Improved
fine-tune with TFM flow instead of standard FM flow (ablation)Wan2.1-T2V-14B · Movie-GenVBench motion score: 0.810 -> 0.857Improved
use residual-type loss instead of solution-type loss (ablation)TFM · VideoJam-BenchVBench motion score: 0.894 -> 0.961Improved
use residual-type loss instead of solution-type loss (ablation)TFM · Movie-GenVBench motion score: 0.811 -> 0.857Improved
apply TFM fine-tuning to Wan2.1 base modelWan2.1-T2V-14B · VideoJam-Bench user study motion% votes favoring TFM: 74.3% preference for TFMImproved
apply TFM fine-tuning to Wan2.1 base modelWan2.1-T2V-14B · Movie-Gen user study motion% votes favoring TFM: 64.6% preference for TFMImproved
apply TFM fine-tuning to Wan2.1 base modelWan2.1-T2V-14B · VideoJam-Bench appearanceVBench appearance score: 0.780 -> 0.783No clear effect
apply TFM fine-tuning to Wan2.1 base modelWan2.1-T2V-14B · Movie-Gen appearanceVBench appearance score: 0.791 -> 0.795No clear effect
fine-tune with TFM instead of FMWan2.1-T2V-1.3B · Temporal perception, frame permutationrelative loss increase after permutationImproved
use TFM instead of FM during trainingWan2.1-T2V-14B · Training step costoverall time per step: 60.51s -> 61.92s (+2.3%)No clear effect
use TFM instead of FM during trainingWan2.1-T2V-1.3B · Training step costoverall time per step: 18.71s -> 19.66s (+5.1%)Hurt
ablated variants without TFM flow or residual loss (ablation)TFM · Movie-Gen / VideoJam motion smoothnessMotion SmoothNo 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