IBM and NASA deploy open-source geospatial AI foundation model on Hugging Face

Hugging Face, the popular platform for open-source models, has just added yet another valuable addition to its repository. IBM and NASA have joined hands to introduce the watsonx.ai geospatial foundation model on Hugging Face, marking a significant milestone in the world of AI and climate science.

The development of this remarkable model was initially revealed back in February, with the primary goal of harnessing the potential of vast volumes of satellite imagery. By leveraging this data, the model aims to contribute to the advancement of climate science and ultimately enhance life on our planet.

The open-source model was meticulously trained on NASA's Harmonized Landsat Sentinel-2 satellite data (HLS) and went through additional fine-tuning using labeled data for various specific use cases, including burn scar and flood mapping. This rigorous training process ensures that the model is well-equipped to provide accurate and valuable insights into critical geospatial applications.

By making this geospatial foundation model available on Hugging Face, IBM and NASA are taking a significant step towards promoting collaboration and knowledge-sharing within the AI community. Researchers, developers, and enthusiasts can now access this cutting-edge model to further their own research, build innovative applications, and make a positive impact on climate science and Earth's well-being.

In conclusion, the introduction of the watsonx.ai geospatial foundation model on Hugging Face is a testament to the power of open-source initiatives and collaborative efforts between leading organizations. It opens up new possibilities for leveraging satellite imagery for climate-related studies and contributes to the collective pursuit of a sustainable future for our planet.