Reproduce “Parametric Trajectory Distillation for Few-Step Video Generation” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Lan Feng, Peter Karkus, Maximilian Igl, Julius Berner, Yuxiao Chen, Shuhan Tan, Alexandre Alahi, Boris Ivanovic and 1 more
PTD distills video generation into four-step students by predicting each denoising segment’s mean velocity and curvature, then training against teacher velocities queried along the student’s predicted path. At inference, curvature is discarded. Across 100 prompts, MiniMax-H3 PTD scored 2.81 versus 1.55 for LightX2V Turbo on seed diversity, while naturalness was similar; this result covers those tested systems and metrics.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.