TVBench

Lost in Time: A New Temporal Benchmark for Video LLMs Daniel Cores , Michael Dorkenwald , Manuel Mucientes, Cees G. M. Snoek, Yuki M. Asano Equal contribution. Updates 23 December 2024: Please redownload the dataset, as the Unexpected Action labels have been updated. TVBench TVBench is a new benchmark specifically created to evaluate temporal understanding in video QA. We identified three main issues in existing datasets: (i) static information from single… See the full description on the dataset page:

Source: Hugging Face Hub (FunAILab/TVBench). Metadata imported from the dataset’s Hub tags.