Machine-readable video datasets optimized for training category-specific large video models. For AI and model teams focused on fashion and retail video understanding, with annotations for trends, product showcases, material textures, and customer interactions.