I have a model that I can load with either torch.load or torch.load_state_dict, but when I do these operations I get back a type that is OrderedDict.
I don’t see where, once a model is trained, it can be loaded a a Module instance and then run to produce some output. In more general terms, I’m looking how to get a Module from the result of a call to one of torch’s load functions, not anything that concerns the shape of the parameters to be returned by an arbitrary model or usage scenario because I know that those are determined by the designers of that model.