VideoKR-Eval

VideoKR-Eval 📄 ArXiv  ｜  💻 Code  ｜  🤗 Collection About This repository contains the VideoKR-Eval benchmark presented in VideoKR: Towards Knowledge- and Reasoning-Intensive Video Understanding (ICML 2026 Spotlight). VideoKR-Eval is an expert-annotated evaluation benchmark for knowledge- and reasoning-intensive video understanding. Unlike existing benchmarks where a substantial fraction of questions can be answered from a single… See the full description on the dataset page:

Source: Hugging Face Hub (minuzero/VideoKR-Eval). Metadata imported from the dataset’s Hub tags.