High-quality Video Caption Dataset | 1,016,722 Sets | Large Language Model(LLM) Data

1,016,722 Sets - High-quality Video Caption Dataset, including 1,016,722 videos and corresponding caption documents in Chinese and English. All videos have various types(scene, architecture, people, plant, animal, unreality, food, item, aerial, time-lapse, slow-motion, close-up), with resolution no less than 1920x1080 and duration no less than 5 seconds. There are two caption languages like Chinese(50 characters or above) and English(22 words or above). This dataset can be used in VLM, video caption and generation tasks. Content 1,016,722 videos, each video has two caption documents in Chinese and English. Format mp4, mov, avi for videos, txt for caption documents. Diversity covering various video content types(scene, architecture, people, plant, animal, unreality, food, item) and video shooting styles(aerial, time-lapse, slow-motion, close-up). Resolution 1,920x1,080 or above. Duration 5s or above. Caption language Chinese(50 characters or above) and English(22 words or above).