Introducing OlmoEarth embeddings: Custom embedding exports from OlmoEarth Studio for downstream analysis

Summary

OlmoEarth Studio now supports custom computation and export of embedding vectors, compact representations of Earth-observation data generated by open-source OlmoEarth foundation models. Users can select an area, a one-to-twelve-month time span, an encoder variant, spatial resolution, and Sentinel-2 L2A or Sentinel-1 RTC imagery, then receive a Cloud-Optimized GeoTIFF with one int8 band per embedding dimension. The post demonstrates similarity search, few-shot segmentation, change detection, and PCA-based exploration; a logistic regression trained on 60 labeled pixels produced a weighted F1 of 0.84 for three land-cover classes in Ca Mau, Vietnam. Embeddings provide a fast, shareable starting point, while supervised fine-tuning can improve performance for applications needing task-specific results; cloud cover, atmospheric artifacts, and missing observations can affect quality.

Takeaways

1. 1

   Studio offers Nano (128-dimensional), Tiny (192-dimensional), and Base (768-dimensional) encoder variants, with 10 m, 20 m, 40 m, or 80 m spatial resolutions and Sentinel-2 L2A, Sentinel-1 RTC, or combined imagery.

2. 2

   A logistic regression using standardized frozen embeddings and only 60 labeled pixels produced a coherent three-class land-cover map for Ca Mau, Vietnam, with weighted F1 of 0.84.

3. 3

   Embedding quality depends partly on input imagery; persistent cloud cover, atmospheric artifacts, and missing observations during the composite period can affect resulting vectors.