Reproduce “StreamEval: Evaluating Interactive Streaming Video Generation Models” 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
StreamEval Team
StreamEval tests interactive video systems across five tasks using timed requests and explicit criteria. It covers 300 scripted sessions lasting 15 to 300 seconds and evaluates 38 models with ten metrics. In reference editing, Vidu S2 Editing scores 15.00 on Edit Success, the highest among four reported systems; automated rubric scores do not establish human preference.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.