Closed Thunfischpirat closed 1 year ago
Currently, the best-performing model in train_model() is saved as a variable best_model=model. Since this isn't a deep copy, continuing training still affects best_model. To fix this, use torch.save().
train_model()
best_model=model
best_model
torch.save()
Currently, the best-performing model in
train_model()is saved as a variablebest_model=model. Since this isn't a deep copy, continuing training still affectsbest_model. To fix this, usetorch.save().