This dataset contains image and video samples of action verbs and verb+noun pairs. It is designed to support machine learning tasks related to visual understanding, action recognition, and language grounding.

Dataset Structure

The dataset contains the following main folders:

action_verbs/images/: image samples organized by action verb categories. action_verbs/videos/: video samples organized by… See the full description on the dataset page: https://huggingface.co/datasets/LSL-datasets/TALKtoME.