Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

OlmoEarth v1.1: A more efficient family of OlmoEarth models

About

We present a set of improvements to the OlmoEarth family. These improvements allow us to cut compute costs during training ($1.7 \times$ reduction in GPU hours required to train our Base models) and inference ($2.9\times$ reductions in MACs on Sentinel-2 tasks), while maintaining the models' overall performance. All training code is available at github.com/allenai/olmoearth_pretrain.

Gabriel Tseng, Yawen Zhang, Favyen Bastani, Henry Herzog, Joseph Redmon, Hadrien Sablon, Piper Wolters, Patrick Alan Johnson, Christopher Wilhelm, Patrick Beukema• 2026

Related benchmarks

TaskDatasetResultRank
Semantic segmentationSen1Floods11 (test)
mIoU80.1
24
Semantic segmentationPASTIS (test)
mIoU30.8
22
Image Classificationm-bigearthnet (test)
µF1 Score72.3
19
Classificationm-so2sat (test)
Mean Accuracy69.8
17
Multi-Label Classificationm-bigearthnet (test)
µF1 Score63.7
4
Time-series classificationCropHarvest PRC (test)
Accuracy82
4
Time-series classificationNandi (test)
Accuracy74.5
4
Object DetectionVessel Detection (test)
F1 Score78.9
2
Semantic segmentationSolar Farm Detection (test)
mIoU84.6
2
Showing 9 of 9 rows

Other info

Follow for update