Dataset · imageomics
TreeOfLife-200M: 233M Wildlife Images
imageomics/TreeOfLife-200M
Giant biological dataset of 233 million photos across 933 thousand taxa for training BioCLIP 2.5-class models.
What's inside
An updated and cleaned version of the largest biodiversity corpus used to train the BioCLIP 2.5 Huge model. Inside are 233 million images covering 933,798 taxa of the tree of life with captions in English and Latin. This revision completes data cleaning and adds previously missed observations from the Observation.org platform.
What you can do with it
- Training specialized vision models for classifying plant, animal, and fungi species
- Creating zero-shot taxon recognition systems for ecological monitoring and citizen science
- Tasks
- Image classification, Zero shot classification
- Languages
- English, Latin
- Size
- 100M–1B rows
- Downloads this month
- 24K
- Likes
- 45
Models trained on it
- imageomics/bioclip-2Zero shot image classification 279.7K
- imageomics/bioclip-2.5-vith14Zero shot image classification 16.6K
- Valen-Peregrine/bioclip-2-tfliteImage feature extraction 88
- Valen-Peregrine/bioclip-2.5-tfliteImage feature extraction 73
- skram/bioclip-2-endpointZero shot image classification 62
- skram/bioclip-2.5-vith14-endpointZero shot image classification 37
- birder-project/vit_l14_pn_bioclip-v2Image feature extraction 33
- birder-project/vit_h14_pn_bioclip-v25Image feature extraction 10
How to load it
from datasets import load_dataset
ds = load_dataset("imageomics/TreeOfLife-200M")