Reusable Deep Learning with EOPF Zarr Samples

Recording

About the webinar

How can we reuse deep learning models trained on EO data and make the most of new-generation datasets? This hands-on webinar shows how to do just that, bridging two powerful tools in the Earth observation ecosystem: the Copernicus Data Space Ecosystem (CDSE) and the EOPF Sentinel Zarr Samples.

The session is part of a wider initiative to support the transition to cloud-native EO data through the EOPF project. Whether you are experimenting with AI models or building EO data pipelines, this webinar helps you unlock reusable deep learning workflows with the latest data formats.

What to expect

The session demonstrates how a deep learning model originally trained on CDSE data using the openEO API can be adapted and reproduced using EOPF Zarr datasets. You will discover:

• Key differences between CDSE and EOPF data access.
• Challenges faced in adapting workflows across ecosystems.
• Practical lessons learned, including code walkthroughs and technical insights.

Who should watch

This webinar is designed for participants with at least an intermediate level of experience working with EO data and machine learning workflows – whether you are part of the research community, developing applications, or just curious about EOPF’s cloud-native potential.

Programme