This can be useful when you have saved model checkpoints during training and want to reload the best checkpoint in the end.
Arguments
- obj
luz object to which you want to copy the new weights.
- path
path to saved model in disk.
- ...
other arguments passed to
torch_load()
.