To get an evaluation of a model version's performances, you need to create a validation set before training a model version.
Once you have trained a model, you will likely want to know how good it performs at the task you are trying to automate.
You can start with a qualitative evaluation here:
You can then access the performance report to get more details on your models:
