MultitaskMultivariateNormal: fix tensor reshape issue

[adamjstewart]

I'm experimenting with a multitask modeling problem and encountered the following error message:

Traceback (most recent call last):
  File "train.py", line 239, in multi
    loss = -criterion(predict_y, train_y)
  File "~/lib/python3.9/site-packages/gpytorch/module.py", line 30, in __call__
    outputs = self.forward(*inputs, **kwargs)
  File "~/lib/python3.9/site-packages/gpytorch/mlls/exact_marginal_log_likelihood.py", line 62, in forward
    res = output.log_prob(target)
  File "~/lib/python3.9/site-packages/gpytorch/distributions/multitask_multivariate_normal.py", line 211, in log_prob
    return super().log_prob(value.view(*value.shape[:-2], -1))
RuntimeError: view size is not compatible with input tensor's size and stride (at least one dimension spans across two contiguous subspaces). Use .reshape(...) instead.

I'm not sure exactly what is causing this issue, but replacing view with reshape solved the issue for me.

Unfortunately I can't share the code or data used to reproduce this issue since it's in-progress research, but I can try to find a minimal reproducible example if you need one.

[gpleiss]

Seems super reasonable - much easier than checking whether the tensor is contiguous. Thanks @adamjstewart !