Tweaaks after re-running expt1

symbolic_nn_tests/dataloader.py

@@ -14,5 +14,11 @@ def create_dataset(
     **kwargs,
 ):
     ds = dataset(DATASET_DIR, download=True, transform=ToTensor())
-    train, val, test = (DataLoader(i, **kwargs) for i in random_split(ds, split))
+    shuffle = kwargs.pop("shuffle", False)
+    shuffle_train = kwargs.pop("shuffle_train", False)
+    to_shuffle = (shuffle or shuffle_train, shuffle, shuffle)
+    train, val, test = (
+        DataLoader(i, shuffle=s, **kwargs)
+        for i, s in zip(random_split(ds, split), to_shuffle)
+    )
     return train, val, test

symbolic_nn_tests/experiment1/model.py

@@ -14,8 +14,8 @@ def collate(batch):
     x, y = zip(*batch)
     x = [i[0] for i in x]
     y = [torch.tensor(i) for i in y]
-    x = torch.stack(x).to("cuda")
+    x = torch.stack(x)
-    y = torch.tensor(y).to("cuda")
+    y = torch.tensor(y)
     return x, y
 
 
@@ -27,14 +27,18 @@ def get_singleton_dataset():
     from symbolic_nn_tests.dataloader import create_dataset
 
     return create_dataset(
-        dataset=QMNIST, collate_fn=collate, batch_size=128, shuffle=True
+        dataset=QMNIST,
+        collate_fn=collate,
+        batch_size=128,
+        shuffle_train=True,
+        num_workers=11,
     )
 
 
 def oh_vs_cat_cross_entropy(y_bin, y_cat):
     return nn.functional.cross_entropy(
         y_bin,
-        nn.functional.one_hot(y_cat),
+        nn.functional.one_hot(y_cat, num_classes=10).float(),
     )
 
 
@@ -56,7 +60,7 @@ def main(
 
     train, val, test = get_singleton_dataset()
     lmodel = TrainingWrapper(
-        model, train_loss=train_loss, val_loss=val_loss, test_loss=val_loss
+        model, train_loss=train_loss, val_loss=val_loss, test_loss=test_loss
     )
     lmodel.configure_optimizers(**kwargs)
     trainer = L.Trainer(max_epochs=20, logger=logger)

symbolic_nn_tests/experiment1/semantic_loss.py

@@ -11,7 +11,9 @@ def create_semantic_cross_entropy(semantic_matrix):
         return ce_loss * semantic_penalty
 
     def oh_vs_cat_semantic_cross_entropy(input_oh, target_cat):
-        return semantic_cross_entropy(input_oh, torch.nn.functional.one_hot(target_cat))
+        return semantic_cross_entropy(
+            input_oh, torch.nn.functional.one_hot(target_cat, num_classes=10).float()
+        )
 
     return oh_vs_cat_semantic_cross_entropy
 

symbolic_nn_tests/experiment2/model.py

@@ -49,7 +49,11 @@ def get_singleton_dataset():
     from symbolic_nn_tests.experiment2.dataset import collate, pubchem
 
     return create_dataset(
-        dataset=pubchem, collate_fn=collate, batch_size=256, shuffle=True
+        dataset=pubchem,
+        collate_fn=collate,
+        batch_size=256,
+        shuffle_train=True,
+        num_workers=11,
     )