default line search to False (can be slow)

cca_zoo/data/deep.py

@@ -9,7 +9,7 @@ class NumpyDataset(Dataset):
     Class that turns numpy arrays into a torch dataset
     """
 
-    def __init__(self, views, labels=None, scale=False, centre=False, precision="float64"):
+    def __init__(self, views, labels=None, scale=False, centre=False, precision="float32"):
         """
 
         :param views: list/tuple of numpy arrays or array likes with the same number of rows (samples)

cca_zoo/models/_stochastic/_base.py

@@ -100,9 +100,9 @@ def fit(self, views: Iterable[np.ndarray], y=None, **kwargs):
 
     def get_dataloader(self, views: Iterable[np.ndarray]):
         if self.batch_size is None:
-            dataset = BatchNumpyDataset(views)
+            dataset = BatchNumpyDataset(views,precision="float64")
         else:
-            dataset = NumpyDataset(views)
+            dataset = NumpyDataset(views,precision="float64")
         if self.val_split is not None:
             train_size = int((1 - self.val_split) * len(dataset))
             val_size = len(dataset) - train_size

cca_zoo/test/test_deepmodels.py

@@ -28,12 +28,12 @@
 Z = rng.rand(256, 14)
 X_conv = rng.rand(256, 1, 16, 16)
 Y_conv = rng.rand(256, 1, 16, 16)
-dataset = NumpyDataset([X, Y, Z], scale=True, centre=True)
+dataset = NumpyDataset([X, Y, Z], scale=True, centre=True, precision="float32")
 check_dataset(dataset)
 train_dataset, val_dataset = random_split(dataset, [200, 56])
 loader = get_dataloaders(dataset)
 train_loader, val_loader = get_dataloaders(train_dataset, val_dataset)
-conv_dataset = NumpyDataset((X_conv, Y_conv))
+conv_dataset = NumpyDataset((X_conv, Y_conv), precision="float32")
 conv_loader = get_dataloaders(conv_dataset)
 train_ids = train_dataset.indices
 epochs = 100