Test performance using Train_generator?

Hi!

I just started week 1 from course 2, and I’m just wondering how to evaluate a model in test set given this train_generator tool.
In the notebooks, they just define train e validation, so they’re used on fit method. But what about evaluate to check test performance and not just see predict results?

Look at the validation_data parameter of the Model#fit method. Return value of ImageDataGenerator#flow_from_directory for the validation data can be used as the parameter value.