Rate this Page
★ ★ ★ ★ ★

torch.autograd.gradcheck.get_numerical_jacobian_wrt_specific_input#

torch.autograd.gradcheck.get_numerical_jacobian_wrt_specific_input(fn, input_idx, inputs, outputs, eps, input=None, is_forward_ad=False)[source]#
Return type:

tuple[Tensor, …]

On this page
Show Source
PyTorch Libraries