wassname/pytorch-ts
1
0
Fork 0
mirror of https://github.com/wassname/pytorch-ts.git synced 2026-08-06 13:30:10 +08:00
Code Issues Packages Projects Releases Wiki Activity

exchange the forward and inverse in Real NVP

Browse Source
This commit is contained in:
Dr. Kashif Rasul
2020-01-22 10:48:45 +01:00
parent f09153cf5a
commit b311aded13
1 changed files with 15 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files
pts/modules/flows.py
+15 -9
View File
@@ -1,5 +1,6 @@
import copy
import math
from abc import ABC, abstractmethod
import torch
import torch.nn as nn
@@ -57,14 +58,14 @@ class FlowSequential(nn.Sequential):
sum_log_abs_det_jacobians = 0
for module in self:
x, log_abs_det_jacobian = module(x, y)
sum_log_abs_det_jacobians = sum_log_abs_det_jacobians + log_abs_det_jacobian
sum_log_abs_det_jacobians += log_abs_det_jacobian
return x, sum_log_abs_det_jacobians
def inverse(self, u, y):
sum_log_abs_det_jacobians = 0
for module in reversed(self):
u, log_abs_det_jacobian = module.inverse(u, y)
sum_log_abs_det_jacobians = sum_log_abs_det_jacobians + log_abs_det_jacobian
sum_log_abs_det_jacobians += log_abs_det_jacobian
return u, sum_log_abs_det_jacobians
@@ -165,12 +166,14 @@ class LinearMaskedCoupling(nn.Module):
# cf RealNVP eq 8 where u corresponds to x (here we're modeling u)
log_s = torch.tanh(s) * (1 - self.mask)
u = (x - t) * torch.exp(-log_s)
u = x * torch.exp(log_s) + t
# u = (x - t) * torch.exp(log_s)
# u = mx + (1 - self.mask) * (x - t) * torch.exp(-s)
# log det du/dx; cf RealNVP 8 and 6; note, sum over input_size done at model log_prob
#log_abs_det_jacobian = -(1 - self.mask) * s
log_abs_det_jacobian = -log_s #.sum(-1, keepdim=True)
# log_abs_det_jacobian = -(1 - self.mask) * s
# log_abs_det_jacobian = -log_s #.sum(-1, keepdim=True)
log_abs_det_jacobian = log_s
return u, log_abs_det_jacobian
@@ -183,11 +186,13 @@ class LinearMaskedCoupling(nn.Module):
t = self.t_net(mu if y is None else torch.cat([y, mu], dim=-1)) * (1 - self.mask)
log_s = torch.tanh(s) * (1 - self.mask)
x = u * torch.exp(log_s) + t
#x = mu + (1 - self.mask) * (u * s.exp() + t) # cf RealNVP eq 7
x = (u - t) * torch.exp(-log_s)
# x = u * torch.exp(log_s) + t
# x = mu + (1 - self.mask) * (u * s.exp() + t) # cf RealNVP eq 7
#log_abs_det_jacobian = (1 - self.mask) * s # log det dx/du
log_abs_det_jacobian = log_s #.sum(-1, keepdim=True)
# log_abs_det_jacobian = (1 - self.mask) * s # log det dx/du
# log_abs_det_jacobian = log_s #.sum(-1, keepdim=True)
log_abs_det_jacobian = -log_s
return x, log_abs_det_jacobian
@@ -363,6 +368,7 @@ class RealNVP(nn.Module):
return sample
class MAF(nn.Module):
def __init__(
self,