Sample data and fine-tuned weights accompanying our ACM SIGSPATIAL 2026 paper, released so
other researchers can run inference with SatMAE, Prithvi, and SpectralGPT on our
multi-temporal crop segmentation and change-detection tasks.
This is not the full training dataset — it's a set of representative chips per region/model
plus the fine-tuned checkpoints, enough to run and sanity-check inference end-to-end.
Contact the authors if you need the complete… See the full description on the dataset page:
https://huggingface.co/datasets/sanmay4119/geofm-agriculture-benchmark.
