PhysAlign: Learning Physical Priors for Dynamical Event-Driven Video Generation via Representation Alignment

Open paper
Benchmark Model Rank Results
text-to-video-generation-on-vbenchPhysAlign (extra data)–VBench Quality Score: 79.05
text-to-video-generation-on-videophyPhysAlign (extra data)–PC: 40.54SA: 53.98

Rank counts only results with a code link.