Dataset · turing-motors
CoVLA driving video dataset with trajectories and descriptions
turing-motors/CoVLA-Dataset
10,000 real driving video clips totaling over 80 hours with trajectory annotations and English text descriptions.
What's inside
A multimodal dataset of video and text for autonomous driving, accepted at WACV 2025 (Oral). It contains 10,000 thirty-second real driving video clips totaling over 80 hours in duration. Each clip includes precise target driving trajectories and detailed English text descriptions of traffic conditions and maneuvers generated through an automated pipeline.
What you can do with it
- Training Vision-Language-Action (VLA) models for autonomous driving
- Training vehicle trajectory prediction systems using video streams
- Training vision-language models (VLM) to describe driving scenarios and maneuvers
- Languages
- English
- Size
- 10K–100K rows
- Downloads this month
- 1.6K
- Likes
- 56
- Access
- on request on Hugging Face
How to load it
from datasets import load_dataset
ds = load_dataset("turing-motors/CoVLA-Dataset")