torch.nn.functional.max_unpool2d
function max_unpool2d(input: Tensor, indices: Tensor, kernel_size: number | [number, number], options?: MaxUnpoolFunctionalOptions): TensorApply 2D max unpooling. Computes the inverse of max_pool2d by scattering values back to their original positions.
Parameters
inputTensor- Input tensor
indicesTensor- Indices of max values
kernel_sizenumber | [number, number]- Size of pooling window
optionsMaxUnpoolFunctionalOptionsoptional- Options for the operation. See
MaxUnpoolFunctionalOptions.