Change kernel_size to self.kernel_size to resolve error in quantized conv module (#24499)
Summary:
Pull Request resolved: https://github.com/pytorch/pytorch/pull/24499
We are attempting to subscript a variable `kernel_size` that may not be
an iterable in the `conv.py` module currently. This leads to errors unless the
user passes in an iterable for kernel_size. D16830855 changed
`self.kernel_size` to be a pair type, but did not actually use the variable.
We want to use `self.kernel_size` which is a pair even if the user passed in an int for `kernel_size` so that we stop having the subscripting error.
Differential Revision: D16859809
fbshipit-source-id: cd2a5497e89d88e518ca7b8a97bf9e69803ee2ba