tfep.utils.math.batch_autograd_log_abs_det_J
- tfep.utils.math.batch_autograd_log_abs_det_J(x, y, **grad_kwargs)[source]
Compute the batch log(abs(det(J))) of
yw.r.t.xusingtorch.autograd.grad.Note that for the determinant to make sense,
xandymust be 2D tensors with identical dimensions.- Parameters:
x (torch.Tensor) – Shape
(batch_size, N). Input.y (torch.Tensor) – Shape
(batch_size, N). Output.**grad_kwargs – Keyword arguments for
autograd.grad(e.g.,create_graph,retain_graph).
- Returns:
log_det_J – Shape
(batch_size,).- Return type:
torch.Tensor