Unable to resume training from check point with custom loss function in NBEATS

[AndresMR77]

Describe the bug

I am training an NBEATS model with a custom loss function defined by myself. I train the model (set save_checkpoints = True) for a certain amount of epochs and then I load it again (load_from_checkpoint) to continue training, but when I do so, the loss function switches to the default MSELoss instead of continuing with my custom loss.

System (please complete the following information):

• Python version: [e.g. 3.7,13]
• darts version [e.g. 0.22.0]