autodiff.OptimizeResult

autodiff.OptimizeResult(params, value, losses, n_steps, converged)

The outcome of :meth:AutodiffPropagation.minimize.

Attributes

Name	Type	Description
params	dict	the optimised parameters, same structure as the ones passed in.
value	float	the loss at params.
losses	np.ndarray	the loss at the starting point and after each update, a 1-D array of length n_steps + 1; its last entry is value.
n_steps	int	the number of updates taken (fewer than requested if gtol was reached).
converged	bool	whether the gradient norm fell below gtol.