Evaluating

Since v.1.0.3.7 there is the possibility to evaluate the models metrics against an evaluation dataset.

 

To do so, there must have a successful training and an evaluation dataset exported with the same format as the Training/ Testing dataset. See Exporting a batch into a ML Dataset.

 

Once we have a good evaluation dataset and the training record for the model, we can click on the "Evaluate" button as shown:

 

 

This will run the Training process in evaluation mode. Once finished the results are saved in the "Eval. folder".