What is `self.loc` in `MixtureCDFFlow` class?

[ovshake]

I am trying to implement the 1D flow as shown in Flow Models Demos (Official).ipynb. I can't figure out what is self.loc in the MixtureCDFFlow class. Also, what does self.n_components denote? This is the code

 def flow(self, x):
        # set up mixture distribution
        weights = F.softmax(self.weight_logits, dim=0).unsqueeze(0).repeat(x.shape[0], 1)
        mixture_dist = self.mixture_dist(self.loc, self.log_scale.exp())
        x_repeat = x.unsqueeze(1).repeat(1, self.n_components)

        # z = cdf of x
        z = (mixture_dist.cdf(x_repeat) * weights).sum(dim=1)

        # log_det = log dz/dx = log pdf(x)
        log_det = (mixture_dist.log_prob(x_repeat).exp() * weights).sum(dim=1).log()

        return z, log_det

Some help is really appreciated.

Thanks

[suryabulusu]

• z = sum over k = 1...3 ( pi_k * normal_cdf (mu_k, sig_k) ) where n_components = 3

• loc and exp(log_scale) are [mu_1, mu_2, mu_3] and [sig_1, sig_2, sig_3] respectively.

• pi_k are the weights of each component.

CDFs map to [0, 1] and we can use uniform distribution for z. Mixture CDFs are preferred as they are more expressive. Hope this helps!