With PR #98 (TF-Record Creation), ML-Enabler will be able to create initial model training data tf-records for a user assuming the user provides a vector of geojson data + imagery.
It would be a nice enhancement to visualize the training data, in a similar manner to how ML-Enabler already has the UI in place to visualize predictions.
It would also be great to be able to visually look at the train, test, and validation split and have those color-coded when the training data is displayed over the imagery on the map.
cc @ingalls
With PR #98 (TF-Record Creation), ML-Enabler will be able to create initial model training data tf-records for a user assuming the user provides a vector of geojson data + imagery.
It would be a nice enhancement to visualize the training data, in a similar manner to how ML-Enabler already has the UI in place to visualize predictions.
It would also be great to be able to visually look at the train, test, and validation split and have those color-coded when the training data is displayed over the imagery on the map.
cc @ingalls