Update kde.py

[ducha-aiki]

[ducha-aiki]

Even with this fix, we have an error:

    y = torch.multinomial(x, min(max_num_cmp,x.shape[-2]), replacement=False)
RuntimeError: cannot sample n_sample > prob_dist.size(-1) samples without replacement

[Parskatt]

I think I was very stupid when I implemented this, let me actually run it for real.

[Parskatt]

I keep confusing the syntax for np.random.choice and torch.multinomial... multinomial gives you inds from some pos measure. Should be more correct now.

[Parskatt]

@ducha-aiki