📚 Documentation

In the master docs for nn.Softmax(), the example provided is:

    >>> m = nn.Softmax()
    >>> input = torch.randn(2, 3)
    >>> output = m(input)

However, this implementation appears deprecated in 1.1. You now get a warning like:

UserWarning: Implicit dimension choice for softmax has been deprecated. Change the call to include dim=X as an argument.

If you agree, I can change the example to m = nn.Softmax(dim=1) and issue a pull request so there is no warning.

Version: I observed the warning with Python 1.1.0 and Python 3.7.1 on OSX.