Code About Equation 6 in the Paper

[xiangyu-sun-789]

Great paper! I love it!

I was looking at the code. I am wondering if z0 = torch.exp(s0) * x0 + t0 and z1 = torch.exp(s) * x1 + t correspond to Equation 6 in the paper. If they do not, could you please let me know which flow class I should use to compute the inverted Equation 6? If they do, shouldn't they be something like z1 = (x1 - t) * torch.exp(-s) instead? By looking at the training method, it looks like the forward method is called for training.

Thank you very much.