Dataset · Vchitect
Video dataset for text-to-video models
Vchitect/Vchitect_T2V_DataVerse
A collection of 14M videos with text descriptions for training text-to-video diffusion models.
What's inside
The dataset includes 14 million high-quality internet videos with detailed text descriptions. It was created by the Shanghai Artificial Intelligence Laboratory team and is used to train the Vchitect-2.0 model.
What you can do with it
- Training and fine-tuning text-to-video diffusion models
- Scaling video generation using Parallel Transformer
- Tasks
- Text → video
- Size
- 1M–10M rows
- Downloads this month
- 23.9K
- Likes
- 11
How to load it
from datasets import load_dataset
ds = load_dataset("Vchitect/Vchitect_T2V_DataVerse")