jvp(func, inputs, v=None, create_graph=False, strict=False)¶
Function that computes the dot product between the Jacobian of the given function at the point given by the inputs and a vector
func (function) – a Python function that takes Tensor inputs and returns a tuple of Tensors or a Tensor.
inputs (tuple of Tensors or Tensor) – inputs to the function
v (tuple of Tensors or Tensor) – The vector for which the Jacobian vector product is computed. Must be the same size as the input of
func. This argument is optional when the input to
funccontains a single element and (if it is not provided) will be set as a Tensor containing a single
create_graph (bool, optional) – If
True, both the output and result will be computed in a differentiable way. Note that when
False, the result can not require gradients or be disconnected from the inputs. Defaults to
strict (bool, optional) – If
True, an error will be raised when we detect that there exists an input such that all the outputs are independent of it. If
False, we return a Tensor of zeros as the jvp for said inputs, which is the expected mathematical value. Defaults to
- tuple with:
func_output (tuple of Tensors or Tensor): output of
jvp (tuple of Tensors or Tensor): result of the dot product with the same shape as the output.
- Return type
>>> def exp_reducer(x): ... return x.exp().sum(dim=1) >>> inputs = torch.rand(4, 4) >>> v = torch.ones(4, 4) >>> jvp(exp_reducer, inputs, v) (tensor([6.3090, 4.6742, 7.9114, 8.2106]), tensor([6.3090, 4.6742, 7.9114, 8.2106]))
>>> jvp(exp_reducer, inputs, v, create_graph=True) (tensor([6.3090, 4.6742, 7.9114, 8.2106], grad_fn=<SumBackward1>), tensor([6.3090, 4.6742, 7.9114, 8.2106], grad_fn=<SqueezeBackward1>))
>>> def adder(x, y): ... return 2 * x + 3 * y >>> inputs = (torch.rand(2), torch.rand(2)) >>> v = (torch.ones(2), torch.ones(2)) >>> jvp(adder, inputs, v) (tensor([2.2399, 2.5005]), tensor([5., 5.]))
The jvp is currently computed by using the backward of the backward (sometimes called the double backwards trick) as we don’t have support for forward mode AD in PyTorch at the moment.