partial unsupervision

model.py

@@ -68,11 +68,12 @@ class ColorTransformerModel(L.LightningModule):
         loss = (1 - alpha) * p_loss + alpha * distance
         # p_loss is unsupervised (preserve relative distances - either in-batch or to-target)
         # distance is supervised.
-        self.log("hp_metric", loss)
+        self.log("hp_metric", distance)
 
         # Log all losses individually
         self.log("train_pres", p_loss)
         self.log("train_mse", distance)
+        self.log("train_loss", loss)
         return loss
 
     def validation_step(self, batch):

newsearch.py

@@ -27,7 +27,7 @@ learning_rate_values = [1e-3]
 # learning_rate_values = [5e-4]
 
 # alpha_values = [0, .25, 0.5, 0.75, 1]  # alpha = 0 is unsupervised. alpha = 1 is supervised.
-alpha_values = [1.0]
+alpha_values = [0.99]
 # widths = [2**k for k in range(4, 13)]
 # depths = [1, 2, 4, 8, 16]
 widths, depths = [512], [1, 2, 4]