Fix np.float and np.float32 to float.

pytorch_forecasting/data/encoders.py

@@ -470,7 +470,7 @@ def _set_parameters(
         if isinstance(y_center, torch.Tensor):
             eps = torch.finfo(y_center.dtype).eps
         else:
-            eps = np.finfo(np.float).eps
+            eps = np.finfo(float).eps
         if self.method == "identity":
             if isinstance(y_center, torch.Tensor):
                 self.center_ = torch.zeros(y_center.size()[:-1])
@@ -785,7 +785,7 @@ def fit(self, y: pd.Series, X: pd.DataFrame):
             self
         """
         y = self.preprocess(y)
-        eps = np.finfo(np.float).eps
+        eps = np.finfo(float).eps
         if len(self.groups) == 0:
             assert not self.scale_by_group, "No groups are defined, i.e. `scale_by_group=[]`"
             if self.method == "standard":

pytorch_forecasting/data/timeseries.py

@@ -1721,12 +1721,12 @@ def _collate_fn(
                     scale = torch.stack([batch[0]["target_scale"][idx] for batch in batches])
                 else:
                     scale = torch.from_numpy(
-                        np.array([batch[0]["target_scale"][idx] for batch in batches], dtype=np.float32),
+                        np.array([batch[0]["target_scale"][idx] for batch in batches], dtype=float),
                     )
                 target_scale.append(scale)
         else:  # convert to tensor
             target_scale = torch.from_numpy(
-                np.array([batch[0]["target_scale"] for batch in batches], dtype=np.float32),
+                np.array([batch[0]["target_scale"] for batch in batches], dtype=float),
             )
 
         # target and weight

pytorch_forecasting/models/nbeats/sub_modules.py

@@ -26,7 +26,7 @@ def linspace(backcast_length: int, forecast_length: int, centered: bool = False)
         norm = backcast_length + forecast_length
         start = 0
         stop = backcast_length + forecast_length - 1
-    lin_space = np.linspace(start / norm, stop / norm, backcast_length + forecast_length, dtype=np.float32)
+    lin_space = np.linspace(start / norm, stop / norm, backcast_length + forecast_length, dtype=float)
     b_ls = lin_space[:backcast_length]
     f_ls = lin_space[backcast_length:]
     return b_ls, f_ls

tests/test_data/test_encoders.py

@@ -163,5 +163,5 @@ def test_TorchNormalizer_dtype_consistency():
     assert TorchNormalizer()(dict(prediction=parameters, target_scale=target_scale)).dtype == torch.float32
     assert TorchNormalizer().transform(parameters, target_scale=target_scale).dtype == torch.float32
 
-    y = np.array([1, 2, 3], dtype=np.float32)
+    y = np.array([1, 2, 3], dtype=float)
     assert TorchNormalizer(method="identity").fit(y).get_parameters().dtype == torch.float32