Databoost/

VidData

	Technical Documentation for the Text-to-Video Dataset “VidData”




	
	1. Introduction

This dataset contains 1006 annotated videos of everyday scenes, used for training and evaluating AI models in video generation and recognition. It is structured to meet the needs of Text-to-Video models and motion analysis.

	2. Dataset Specifications




	
	2.1. Generation Criteria

Maximum video duration: 10 seconds maximum Video themes: Walking Exercising Writing Shopping… See the full description on the dataset page: https://huggingface.co/datasets/Databoost/VidData.