Dataset · project-oceania
17 million labeled plankton images
project-oceania/planktonzilla-17M
Largest combined dataset of plankton photographs with taxonomic labeling.
What's inside
The dataset combines 17 million plankton images from all available open sources with annotations. Each image is provided with a standardized taxonomic hierarchy in English. The collection covers various imaging systems and oceanographic environments.
What you can do with it
- Training models for plankton species classification and recognition
- Training multimodal models (image-to-text) for scientific analytics
- Generating synthetic microorganism images
- Tasks
- Image → text, Image classification, Image text → text, Image text → image
- Languages
- English
- Size
- 10M–100M rows
- Downloads this month
- 1.7K
- Likes
- 8
Models trained on it
- project-oceania/timm-eva02-large-m38m-ft-planktonzillaImage classification 2
- project-oceania/CLIP-planktonzilla-epoch30Zero shot image classification 0
- project-oceania/CLIP-ViT-L-14.laion2b-pt.planktonzilla-ptZero shot image classification 0
- project-oceania/CLIP-ViT-L-14.bioclip2-pt.planktonzilla-ptZero shot image classification 0
- project-oceania/CLIP-ViT-B-16.bioclip-pt.planktonzilla-ptZero shot image classification 0
- project-oceania/CLIP-ViT-B-16.openai-pt.planktonzilla-ptZero shot image classification 0
How to load it
from datasets import load_dataset
ds = load_dataset("project-oceania/planktonzilla-17M")