torch.optim.functional.rprop

torch.optim.functional.rprop(params, grads, prevs, step_sizes, state_steps, foreach=None, capturable=False, maximize=False, differentiable=False, has_complex=False, *, step_size_min, step_size_max, etaminus, etaplus)

Functional API that performs Rprop algorithm computation.

This function updates the provided parameters and optimizer state in place. The caller must initialize and retain optimizer state. Unless intentionally constructing a differentiable update with a supported differentiable=True argument, call this function under torch.no_grad. See Functional optimizer API for the common functional optimizer contract and examples, and Rprop for algorithm details.