add new vae and small changes in basic code

demo/vae_gaussian_bernoulli.py

@@ -41,11 +41,6 @@
     datasets.MNIST('./data', train=False, transform=transforms.ToTensor()),
     batch_size=args.batch_size, shuffle=True, **kwargs)
 
-INITIAL_TEMP = 1.0
-ANNEAL_RATE = 0.00003
-MIN_TEMP = 0.1
-
-temp = INITIAL_TEMP
 steps = 0
 
 
@@ -66,7 +61,7 @@ def decode(self, z):
         h3 = F.relu(self.fc3(z))
         return torch.sigmoid(self.fc4(h3))
 
-    def forward(self, x, temp=1.0, hard=False):
+    def forward(self, x, hard=False):
         mu = self.encode(x.view(-1, 784))
         q_z = GaussianRelaxedBernoulli(mu, torch.tensor([1], device=device))
         z = q_z.rsample()  # sample with reparameterization
@@ -84,26 +79,26 @@ def forward(self, x, temp=1.0, hard=False):
 
 
 # Reconstruction + KL divergence losses summed over all elements and batch
-def loss_function(recon_x, x, mu, prior=0.5, eps=1e-10):
+def loss_function(recon_x, x, q_z, prior=0.5, eps=1e-10):
     BCE = F.binary_cross_entropy(recon_x, x.view(-1, 784), reduction='sum')
     # You can also compute p(x|z) as below, for binary output it reduces
     # to binary cross entropy error, for gaussian output it reduces to
-    t1 = mu * ((mu + eps) / prior).log()
-    t2 = (1 - mu) * ((1 - mu + eps) / (1 - prior)).log()
+    t1 = q_z * ((q_z + eps) / prior).log()
+    t2 = (1 - q_z) * ((1 - q_z + eps) / (1 - prior)).log()
     KLD = torch.sum(t1 + t2, dim=-1).sum()
 
     return BCE + KLD
 
 
 def train(epoch):
-    global temp, steps
+    global steps
     model.train()
     train_loss = 0
     for batch_idx, (data, _) in enumerate(train_loader):
         data = data.to(device)
         optimizer.zero_grad()
-        recon_batch, q_z = model(data, temp=temp)
-        loss = loss_function(recon_batch, data, q_z)
+        recon_batch, z = model(data)
+        loss = loss_function(recon_batch, data, z)
         loss.backward()
         train_loss += loss.item()
         optimizer.step()
@@ -115,22 +110,19 @@ def train(epoch):
                 loss.item() / len(data)))
 
         steps += 1
-        if steps % 1000 == 0:
-            temp = max(temp * np.exp(-ANNEAL_RATE * steps), MIN_TEMP)
 
     print('====> Epoch: {} Average loss: {:.4f}'.format(
           epoch, train_loss / len(train_loader.dataset)))
 
 
 def test(epoch):
-    global temp
     model.eval()
     test_loss = 0
     with torch.no_grad():
         for i, (data, _) in enumerate(test_loader):
             data = data.to(device)
-            recon_batch, q_z = model(data, temp=temp)
-            test_loss += loss_function(recon_batch, data, q_z).item()
+            recon_batch, z = model(data)
+            test_loss += loss_function(recon_batch, data, z).item()
             if i == 0:
                 n = min(data.size(0), 8)
                 comparison = torch.cat([data[:n],

demo/vae_straight_through_bernoulli.py

@@ -0,0 +1,149 @@
+import os
+import argparse
+import numpy as np
+import torch
+import sys
+import torch.utils.data
+from torch import nn, optim
+from torch.nn import functional as F
+from torchvision import datasets, transforms
+from torchvision.utils import save_image
+
+sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), '..', 'src')))
+from relaxit.distributions import StraightThroughBernoulli
+
+parser = argparse.ArgumentParser(description='VAE MNIST Example')
+parser.add_argument('--batch-size', type=int, default=128, metavar='N',
+                    help='input batch size for training (default: 128)')
+parser.add_argument('--epochs', type=int, default=10, metavar='N',
+                    help='number of epochs to train (default: 10)')
+parser.add_argument('--no-cuda', action='store_true', default=False,
+                    help='enables CUDA training')
+parser.add_argument('--seed', type=int, default=1, metavar='S',
+                    help='random seed (default: 1)')
+parser.add_argument('--log_interval', type=int, default=10, metavar='N',
+                    help='how many batches to wait before logging training status')
+args = parser.parse_args()
+args.cuda = not args.no_cuda and torch.cuda.is_available()
+
+torch.manual_seed(args.seed)
+
+device = torch.device("cuda" if args.cuda else "cpu")
+
+os.makedirs('./results/vae_straight_through_bernoulli', exist_ok=True)
+
+kwargs = {'num_workers': 1, 'pin_memory': True} if args.cuda else {}
+train_loader = torch.utils.data.DataLoader(
+    datasets.MNIST('./data', train=True, download=True,
+                   transform=transforms.ToTensor()),
+    batch_size=args.batch_size, shuffle=True, **kwargs)
+test_loader = torch.utils.data.DataLoader(
+    datasets.MNIST('./data', train=False, transform=transforms.ToTensor()),
+    batch_size=args.batch_size, shuffle=True, **kwargs)
+
+steps = 0
+
+
+class VAE(nn.Module):
+    def __init__(self):
+        super(VAE, self).__init__()
+
+        self.fc1 = nn.Linear(784, 400)
+        self.fc2 = nn.Linear(400, 20)
+        self.fc3 = nn.Linear(20, 400)
+        self.fc4 = nn.Linear(400, 784)
+
+    def encode(self, x):
+        h1 = F.relu(self.fc1(x))
+        return self.fc2(h1)
+
+    def decode(self, z):
+        h3 = F.relu(self.fc3(z))
+        return torch.sigmoid(self.fc4(h3))
+
+    def forward(self, x, hard=False):
+        a = self.encode(x.view(-1, 784))
+        q_z = StraightThroughBernoulli(a)
+        z = q_z.rsample()  # sample with reparameterization
+
+        if hard:
+            # No step function in torch, so using sign instead
+            z_hard = 0.5 * (torch.sign(z) + 1)
+            z = z + (z_hard - z).detach()
+
+        return self.decode(z), a
+
+
+model = VAE().to(device)
+optimizer = optim.Adam(model.parameters(), lr=1e-3)
+
+
+# Reconstruction + KL divergence losses summed over all elements and batch
+def loss_function(recon_x, x, a, prior=0.5, eps=1e-10):
+    BCE = F.binary_cross_entropy(recon_x, x.view(-1, 784), reduction='sum')
+    # You can also compute p(x|z) as below, for binary output it reduces
+    # to binary cross entropy error, for gaussian output it reduces to
+
+    # p = вер выпадения 1 в straight through
+    p = torch.nn.functional.sigmoid(a)
+    t1 = p * ((p + eps) / prior).log()
+    t2 = (1 - p) * ((1 - p + eps) / (1 - prior)).log()
+    KLD = torch.sum(t1 + t2, dim=-1).sum() 
+
+    return BCE + KLD
+
+
+def train(epoch):
+    global steps
+    model.train()
+    train_loss = 0
+    for batch_idx, (data, _) in enumerate(train_loader):
+        data = data.to(device)
+        optimizer.zero_grad()
+        recon_batch, a = model(data)
+        loss = loss_function(recon_batch, data, a)
+        loss.backward()
+        train_loss += loss.item()
+        optimizer.step()
+
+        if batch_idx % args.log_interval == 0:
+            print('Train Epoch: {} [{}/{} ({:.0f}%)]\tLoss: {:.6f}'.format(
+                epoch, batch_idx * len(data), len(train_loader.dataset),
+                100. * batch_idx / len(train_loader),
+                loss.item() / len(data)))
+
+        steps += 1
+
+    print('====> Epoch: {} Average loss: {:.4f}'.format(
+          epoch, train_loss / len(train_loader.dataset)))
+
+
+def test(epoch):
+    model.eval()
+    test_loss = 0
+    with torch.no_grad():
+        for i, (data, _) in enumerate(test_loader):
+            data = data.to(device)
+            recon_batch, a = model(data)
+            test_loss += loss_function(recon_batch, data, a).item()
+            if i == 0:
+                n = min(data.size(0), 8)
+                comparison = torch.cat([data[:n],
+                                       recon_batch.view(args.batch_size, 1, 28, 28)[:n]])
+                save_image(comparison.cpu(),
+                           'results/vae_straight_through_bernoulli/reconstruction_' + str(epoch) + '.png', nrow=n)
+
+    test_loss /= len(test_loader.dataset)
+    print('====> Test set loss: {:.4f}'.format(test_loss))
+
+
+if __name__ == "__main__":
+    for epoch in range(1, args.epochs + 1):
+        train(epoch)
+        test(epoch)
+        with torch.no_grad():
+            sample = np.random.binomial(1, 0.5, size=(64, 20))
+            sample = torch.from_numpy(np.float32(sample)).to(device)
+            sample = model.decode(sample).cpu()
+            save_image(sample.view(64, 1, 28, 28),
+                       'results/vae_straight_through_bernoulli/sample_' + str(epoch) + '.png')