Doesn't work on PyTorch 1.8

[interfect]

PyTorch recently had a 1.8 release, bringing much better support for backing torch.cuda tensors with AMD GPUs.

However, CLIP hasn't been ported to PyTorch 1.8 yet.

In particular, CLIP still uses the deprecated and now removed pytorch.irfft, which needs to be replaced with calls to methods in the torch.fft namespace to work on PyTorch 1.8.

Unfortunately, the PR that removed support for the old methods doesn't provide a recipe for translating calls that can be executed by someone who doesn't understand the finer points of FFTs. It seems to me that the square-root-of-a-bunch-of-stuff normalization method of the old function isn't available as any of the normalization modes of torch.fft.irfft, and I'm not sure of the number of dimensions involved here, or whether we have the input versus the output sizes handy.

[interfect]

Sorry, this is on the wrong project.

[jongwook]

FWIW, CLIP doesn't work on 1.8 either. I'm keeping #49 open to track the compatibility fix.