Skip to content

Latest commit

 

History

History
14 lines (9 loc) · 707 Bytes

README.md

File metadata and controls

14 lines (9 loc) · 707 Bytes

SatU-net

banner

How to Use

  • Download the files to a seperate folder on your machine. Unzip the fields model. Download and unzip the classifier if you are not training your own.

  • Follow the instructions in the respective notebooks

  • 10 sample images are provided to test the code and model predictions

Update

Implemented a vectorized one-hot maker into the training itself. This negates the need to save one-hot files to disk. This method reads files at 1/3rd the speed as loading the array from disk. However, this should speed up with lower resolution images