var_mean¶

dragon.vm.torch.var_mean(
  input,
  dim=None,
  keepdim=False,
  out=None
)[source]¶

Compute the variance and mean of elements along the given dimension.

dim could be negative or None:

x = torch.tensor([[1, 2, 3], [4, 5, 6]], dtype=torch.float32)

# A negative dimension is the last-k dimension
print(torch.var_mean(x, dim=1))
print(torch.var_mean(x, dim=-1))  # Equivalent

# If dimension is None, reduce input as a vector
# and return a scalar result
print(torch.var_mean(x))  # (2.917, 3.5)

# Also, dimension could be a sequence of integers
print(torch.var_mean(x, (0, 1)))  # (2.917, 3.5)

Parameters:

• input (dragon.vm.torch.Tensor) – The input tensor.
• dim (Union[int, Sequence[int]], optional) – The dimension to reduce.
• keepdim (bool, optional, default=False) – Keep the reduced dimension or not.
• out (Sequence[dragon.vm.torch.Tensor], optional) – The optional output value and index.

Returns:

Sequence[dragon.vm.torch.Tensor] – The variance and mean tensor.