Some detailed problem about torch.load_state_dict()

[Amrit_Das]

How to load part of pre trained model?

I’m afraid not The keys of state_dict must exactly match the keys returned by this module’s state_dict() function. if name not in new model , it will raise KeyError but I gusee this may work for you def load_my_state_dict(self, state_dict): own_state = self.state_dict() for name, param in state_dict.items(): if name not in own_state: continue if isinstance(param, Parameter): # backwards compatibility for seria…

This is a similiar kind of answer from

How to load part of pre trained model?

How to load part of pre trained model? Will the unused parameters be auto deleted if I load the whole model but only use part of it?