Hello,
I need to train one model and copy its weights to new model. Then I want to train the new model without training the old model.
Does doing deep copying will help?
Hello,
I need to train one model and copy its weights to new model. Then I want to train the new model without training the old model.
Does doing deep copying will help?
A deep copy should work. Alternatively, you could store the state_dict
and reload it for your new model.