Dataset · madebyollin
10 million links to free Flickr photos
madebyollin/megalith-10m
Large set of Flickr image links in public domain and CC0 for training.
What's inside
Megalith-10m contains about 10 million Flickr image links tagged as photo with Public Domain, CC0, U.S. Government Work, or No known copyright restrictions licenses. As intended by the author, these are only links to wholesome unedited uncopyrighted photographs; images must be downloaded via links. Modality is image+text, exact metadata format in the card is not disclosed.
What you can do with it
- Collection of training corpus for text-to-image
- Image filtering and deduplication by license
- Research on license-clean datasets
- Size
- 1M–10M rows
- Downloads this month
- 1.1K
- Likes
- 108
Models trained on it
- aipicasso/commonart-betaText → image 158
- prithivMLmods/OpenSDI-Flux.1-SigLIP2Image classification 77
- prithivMLmods/OpenSDI-SD1.5-SigLIP2Image classification 23
- prithivMLmods/OpenSDI-SD2.1-SigLIP2Image classification 18
- prithivMLmods/OpenSDI-SD3-SigLIP2Image classification 14
- prithivMLmods/OpenSDI-SDXL-SigLIP2Image classification 14
How to load it
from datasets import load_dataset
ds = load_dataset("madebyollin/megalith-10m")