Update scGen keras network

theislab · Aug 21, 2019 · 42f7a92 · 42f7a92
1 parent 22dcc45
commit 42f7a92
Showing 1 changed file with 6 additions and 11 deletions.
diff --git a/scgen/models/_vae_keras.py b/scgen/models/_vae_keras.py
@@ -98,7 +98,7 @@ def _encoder(self):
         log_var = Dense(self.z_dim, kernel_initializer=self.init_w)(h)
         z = Lambda(self._sample_z, output_shape=(self.z_dim,), name="Z")([mean, log_var])
 
-        self.encoder_model = Model(inputs=self.x, outputs=[mean, log_var, z], name="encoder")
+        self.encoder_model = Model(inputs=self.x, outputs=z, name="encoder")
         return mean, log_var
 
     def _decoder(self):
@@ -178,7 +178,7 @@ def _create_network(self):
         self.mu, self.log_var = self._encoder()
 
         self.x_hat = self._decoder()
-        self.vae_model = Model(inputs=self.x, outputs=self.decoder_model(self.encoder_model(self.x)[2]), name="VAE")
+        self.vae_model = Model(inputs=self.x, outputs=self.decoder_model(self.encoder_model(self.x)), name="VAE")
 
     def _loss_function(self):
         """
@@ -224,7 +224,7 @@ def to_latent(self, data):
             latent: numpy nd-array
                 Returns array containing latent space encoding of 'data'
         """
-        latent = self.encoder_model.predict(data)[2]
+        latent = self.encoder_model.predict(data)
         return latent
 
     def _avg_vector(self, data):
@@ -246,7 +246,7 @@ def _avg_vector(self, data):
         latent_avg = numpy.average(latent, axis=0)
         return latent_avg
 
-    def reconstruct(self, data, use_data=False):
+    def reconstruct(self, data):
         """
             Map back the latent space encoding via the decoder.
 
@@ -265,11 +265,6 @@ def reconstruct(self, data, use_data=False):
             rec_data: 'numpy nd-array'
                 Returns 'numpy nd-array` containing reconstructed 'data' in shape [n_obs, n_vars].
         """
-        # if use_data:
-        #     latent = data
-        # else:
-        #     latent = self.to_latent(data)
-        # rec_data = self.sess.run(self.x_hat, feed_dict={self.z_mean: latent, self.is_training: False})
         rec_data = self.decoder_model.predict(x=data)
         return rec_data
 
@@ -321,7 +316,7 @@ def linear_interpolation(self, source_adata, dest_adata, n_steps):
             vector = start * (1 - alpha) + end * alpha
             vectors[i, :] = vector
         vectors = numpy.array(vectors)
-        interpolation = self.reconstruct(vectors, use_data=True)
+        interpolation = self.reconstruct(vectors)
         return interpolation
 
     def predict(self, adata, conditions, cell_type_key, condition_key, adata_to_predict=None, celltype_to_predict=None, obs_key="all"):
@@ -393,7 +388,7 @@ def predict(self, adata, conditions, cell_type_key, condition_key, adata_to_pred
         else:
             latent_cd = self.to_latent(ctrl_pred.X)
         stim_pred = delta + latent_cd
-        predicted_cells = self.reconstruct(stim_pred, use_data=True)
+        predicted_cells = self.reconstruct(stim_pred)
         return predicted_cells, delta
 
     def restore_model(self):