athena.nll.BackwardNet.forward¶
-
BackwardNet.forward(mapped_inputs)[source]¶ Maps transformed inputs to original inputs.
- Parameters
mapped_inputs (torch.Tensor) – DoubleTensor n_samples-by-n_params containing the nonlinear transformed inputs.
- Return inputs
DoubleTensor n_samples-by-n_params with the points in the original full input space.
- Return type