How to export the model with best weights

[CarlosNacher]

Hello!

One question, perhaps is a Pytorch general doubt, when you model.to_onnx (or to_torch for example), the model you are exporting does use the best weights or the last ones? And when in Anomalib we use the export method of the Engine class, what happens? If the last ones, how should we act to export the best weights taking advantage of the Anomalib's methods?

Thank you so much.