Reproduce “Video-Coder: Reinforcing Agentic Coding for 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
Chaoyang Wang, Sicheng Gao, Junan Hu, Kunyang Li, Dongyang Chen, Yiyang Zhou, Huaxiu Yao, Yong Jae Lee and 2 more
Video-Coder uses a coding agent to build and render an executable 3D scene, then guides a video generator with the draft. The agent can inspect execution feedback and frames and revise its code. With MiniMax-H3, it scores 0.67 versus 0.61 on PhyGenBench and 78.19% versus 54.58% on VBench-2.0 physical dimensions; its material score remains below CogVideoX-1.5’s 83.19%. These benchmark results cover tested settings.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.