SoilNet pretrained model v1.0.1

Pretrained model tarball for inference and fine tuning.

Release Version v2.0.1

This is the code for the IEEE TGRS published version of the paper titled: "SSL-SoilNet: A Hybrid Transformer-based Framework with Self-Supervised Learning for Large-scale Soil Organic Carbon Prediction"

Version 2.0.1 changes:

• Added .py files with command line options and flags for easy use of the model.
• Removed Notebook files; you still can access them in Release 2.0.0

Version 2.0.0 changes:

• Added self-supervised contrastive learning for multimodal data
• Added a second dataset analysis titled 'RaCA.'
• More ML models were added for comparison.

Release Version v2.0.0

This is the code for the new version of the paper titled: "SSCL-SoilNet: A Hybrid Deep Learning Framework with Self-Supervised Contrastive Learning for Large-scale Soil Organic Carbon Prediction"

• Added self-supervised contrastive learning for multimodal data
• Added a second dataset analysis titled 'RaCA.'
• More ML models were added for comparison.

Review Code at : notebook version

Preprint Version

This is the code for the preprint version of this paper titled: "SoilNet: An Attention-based Spatio-temporal Deep Learning Framework for Soil Organic Carbon Prediction with Digital Soil Mapping in Europe"

• No Self-Supervision was used in this model.
• It is mainly CNN/LSTM based and no transformer is used (expect for the attention module).

Review Code at : SoilNet/tree/preprint_version