Multi-Modal Foundation · allenai
OlmoEarth v1 FT — Ecosystem Type Mapping (Base)
Predicts ecosystem categories from satellite imagery across North Africa for biodiversity and conservation workflows.
What OlmoEarth v1 FT — Ecosystem Type Mapping (Base) is
OlmoEarth v1 FT — Ecosystem Type Mapping (Base) is an Earth observation model developed by the Allen Institute for AI. Fine-tuned from OlmoEarth-v1-Base, it is designed to predict ecosystem categories directly from satellite imagery to support ecological research and environmental monitoring.
The model addresses the challenge of identifying and mapping ecosystem distributions for conservation, climate, and biodiversity programmes. It classifies landscapes using categories defined by the IUCN Global Ecosystem Typology, providing standardized ecological outputs across the North Africa region.
How it works
The model builds upon the OlmoEarth-v1-Base architecture through fine-tuning on expert-annotated labels supplied by the Global Ecosystem Atlas. The training data focuses specifically on the North Africa region, aligning satellite image inputs to corresponding categories in the IUCN Global Ecosystem Typology.
Key facts
- Base model
- OlmoEarth-v1-Base
- Domain
- Earth observation, conservation, and climate
- Target region
- North Africa
- Taxonomy
- IUCN Global Ecosystem Typology
- Data provider
- Global Ecosystem Atlas
What you can use it for
- Ecosystem type classification
- Classifies satellite imagery across North Africa into standardized ecological categories.
- Conservation planning
- Delineates ecosystem boundaries to assist conservation teams in monitoring ecological changes and habitat distribution.
- Climate and biodiversity analysis
- Supplies structured ecosystem mapping data for environmental assessments and biodiversity research.
Run it here
We list this model but cannot run it yet: no pipeline_tag — needs a human.
Everything on this page still applies. Running it needs a loader for this repository, not a bigger machine.
Strengths and limits
Good at
- Trained on expert-annotated ground-truth data from the Global Ecosystem Atlas.
- Outputs correspond directly to the standardized IUCN Global Ecosystem Typology.
- Tailored specifically for ecological and geographical conditions in North Africa.
Watch out for
- Trained specifically for North Africa, limiting direct application to other geographical regions.
- Underlying training data from the Global Ecosystem Atlas is scheduled for release in 2026.
Inputs and outputs
You provide
Sequence or text
A protein or nucleotide sequence, or free text — whichever this model was trained on.
You get back
Mean-pooled representation, one vector per input.
Licence
Check the licence before commercial use · See model card
The See model card licence carries terms worth reading before you depend on it.
Questions
Can I use OlmoEarth v1 FT — Ecosystem Type Mapping (Base) commercially?
Check the licence before commercial use. The See model card licence carries terms worth reading before you depend on it.
What ecosystem classification scheme does this model use?
Predictions follow the categories defined in the IUCN Global Ecosystem Typology.
Which geographical areas does the model cover?
The model is trained specifically for the North Africa region.
What is the base model for this fine-tune?
It is fine-tuned from OlmoEarth-v1-Base.
Who provided the training annotations?
The training data consists of expert annotations provided by the Global Ecosystem Atlas.
Written from OlmoEarth v1 FT — Ecosystem Type Mapping (Base)'s own documentation (model card). Statements citing a figure that does not appear there are removed before publication. Licence and running details come from our registry.