fix (ctgan): add discriminator to model attributes

sdv-dev · Nov 21, 2024 · fc0b3e3 · fc0b3e3
1 parent c0ea824
commit fc0b3e3
Showing 1 changed file with 10 additions and 7 deletions.
diff --git a/ctgan/synthesizers/ctgan.py b/ctgan/synthesizers/ctgan.py
@@ -190,6 +190,7 @@ def __init__(
         self._transformer = None
         self._data_sampler = None
         self._generator = None
+        self._discriminator = None
         self.loss_values = None
 
     @staticmethod
@@ -330,7 +331,7 @@ def fit(self, train_data, discrete_columns=(), epochs=None):
             self._embedding_dim + self._data_sampler.dim_cond_vec(), self._generator_dim, data_dim
         ).to(self._device)
 
-        discriminator = Discriminator(
+        self._discriminator = Discriminator(
             data_dim + self._data_sampler.dim_cond_vec(), self._discriminator_dim, pac=self.pac
         ).to(self._device)
 
@@ -342,7 +343,7 @@ def fit(self, train_data, discrete_columns=(), epochs=None):
         )
 
         optimizerD = optim.Adam(
-            discriminator.parameters(),
+            self._discriminator.parameters(),
             lr=self._discriminator_lr,
             betas=(0.5, 0.9),
             weight_decay=self._discriminator_decay,
@@ -395,10 +396,10 @@ def fit(self, train_data, discrete_columns=(), epochs=None):
                         real_cat = real
                         fake_cat = fakeact
 
-                    y_fake = discriminator(fake_cat)
-                    y_real = discriminator(real_cat)
+                    y_fake = self._discriminator(fake_cat)
+                    y_real = self._discriminator(real_cat)
 
-                    pen = discriminator.calc_gradient_penalty(
+                    pen = self._discriminator.calc_gradient_penalty(
                         real_cat, fake_cat, self._device, self.pac
                     )
                     loss_d = -(torch.mean(y_real) - torch.mean(y_fake))
@@ -423,9 +424,9 @@ def fit(self, train_data, discrete_columns=(), epochs=None):
                 fakeact = self._apply_activate(fake)
 
                 if c1 is not None:
-                    y_fake = discriminator(torch.cat([fakeact, c1], dim=1))
+                    y_fake = self._discriminator(torch.cat([fakeact, c1], dim=1))
                 else:
-                    y_fake = discriminator(fakeact)
+                    y_fake = self._discriminator(fakeact)
 
                 if condvec is None:
                     cross_entropy = 0
@@ -520,3 +521,5 @@ def set_device(self, device):
         self._device = device
         if self._generator is not None:
             self._generator.to(self._device)
+        if self._discriminator is not None:
+            self._discriminator.to(self._device)