nn.functional.upsample `scale_factor` parameter description makes no sense

[zou3519]

📚 Documentation

https://pytorch.org/docs/master/nn.functional.html#torch.nn.functional.upsample

It says it "has to be an integer", but the parenthesis says it can be a "float or Tuple[Float]".