Add option to Trainer to save best CALIBRATED model

neurostatslab / vocalocator

Deep neural networks for sound source localization and vocalization attribution.

MIT License

2 stars 0 forks source link

Closed amanchoudhri closed 3 months ago

amanchoudhri commented 1 year ago

Also potentially track negative log likelihood on validation set in addition to error in centimeters. Allowing user to specify which models to save?