What is the difference between state_dict and model.parameters() in PyTorch?

state_dict stores named tensors of model weights, while parameters() returns iterable parameter objects for optimization.

state_dict is a Python dictionary mapping parameter/buffer names to tensors, making it ideal for saving/loading models. parameters() returns raw parameter tensors used by optimizers. Buffers like running_mean in BatchNorm are included in state_dict but not parameters().