Default MvNormal covariance doesn't make sense

[ricardoV94]

pytensor/pytensor/tensor/random/basic.py

Lines 897 to 903 in 5d4b0c4

	dtype = pytensor.config.floatX if self.dtype == "floatX" else self.dtype
	if mean is None:
	mean = np.array([0.0], dtype=dtype)
	if cov is None:
	cov = np.array([[1.0]], dtype=dtype)
	return super().__call__(mean, cov, size=size, **kwargs)

It is not allowed to broadcast with mu so it should be aware of its shape. And should probably be eye(mu.shape[-1]), not ones. Or we can just remove it, since numpy does not provide a default covariance either?

[aseyboldt]

I don't think we should have a default covariance. If anything, that should be the identity, but if that's what you want, why would you use MvNormal and not Normal?

[ricardoV94]

So let's take away the default