when saving an optimizier during training with optimizer.state_dict()
and later loading it via torch.load() I get a dictionary containing “state” and “param_groups”
as described in the documentation https://pytorch.org/docs/stable/optim.html?highlight=state_dict#torch.optim.Optimizer.load_state_dict.
In “params_group” I have an entry for “params”: what exactly are these values?
Are these the last gradient values that have been used to update my NN?
Thanks in advance!