How did having 2 extra bits of resolution in OLI2 vs OLI (Operational Land Imager) affect the training process of the foundation model?

#16
by hyzhak - opened

How did having 2 extra bits of resolution in OLI2 vs OLI (Operational Land Imager) affect the training process of the foundation model?
Note: Based on https://www.earthdata.nasa.gov/sensors/oli OLI2 has 14bits vs 12bits for OLI.

IBM NASA Geospatial org

For finetuning, this should not make any difference. The model will adapt to the type of data you will provide. If there are major differences, it make take a little bit longer to converge.

Sign up or log in to comment