Dataset · gzqy1026
Satellite images with text descriptions
gzqy1026/RemoteCLIP
A large corpus of image-text pairs for Earth remote sensing tasks in English.
What's inside
The dataset of 100K to 1M examples contains raw image bytes (e.g., airport images from RSICD) and English text descriptions. The structure includes a CSV file with header and filename columns, with repeating rows visible in the card preview.
What you can do with it
- Training multimodal CLIP-type models for satellite images
- Fine-tuning image-to-text models to generate aerial photography descriptions
- Training text-to-image generators for top-down view synthesis
- Tasks
- Image → text, Text → image
- Languages
- English
- Size
- 100K–1M rows
- Downloads this month
- 242
- Likes
- 11
How to load it
from datasets import load_dataset
ds = load_dataset("gzqy1026/RemoteCLIP")