fix typy hint noise_range_t

timm/scheduler/cosine_lr.py

@@ -26,23 +26,23 @@ class CosineLRScheduler(Scheduler):
     k-decay option based on `k-decay: A New Method For Learning Rate Schedule` - https://arxiv.org/abs/2004.05909
 
     Args:
-        optimizer (torch.optim.Optimizer): torch optimizer to schedule
+        optimizer (torch.optim.Optimizer): torch optimizer to schedule  
-        t_initial (int): Number of epochs it initial (first) cycle.
+        t_initial (int): Number of epochs it initial (first) cycle.  
-        lr_min (float, optional):  Minimum learning rate to use during the scheduling. Defaults to 0..
+        lr_min (float, optional):  Minimum learning rate to use during the scheduling. Defaults to 0..  
-        cycle_mul (float, optional): Multiplyer for cycle length. Defaults to 1..
+        cycle_mul (float, optional): Multiplyer for cycle length. Defaults to 1..  
-        cycle_decay (float, optional): Factor to decay lr at next cycle. Defaults to 1..
+        cycle_decay (float, optional): Factor to decay lr at next cycle. Defaults to 1..  
-        cycle_limit (int, optional): Number of cycles. Defaults to 1.
+        cycle_limit (int, optional): Number of cycles. Defaults to 1.  
-        warmup_t (int, optional): Number of epochs to warmup. Defaults to 0.
+        warmup_t (int, optional): Number of epochs to warmup. Defaults to 0.  
-        warmup_lr_init (float, optional): Initial learning rate during warmup . Defaults to 0.
+        warmup_lr_init (float, optional): Initial learning rate during warmup . Defaults to 0.  
-        warmup_prefix (bool, optional): If True, after warmup annealing starts from initial LR. Defaults to False.
+        warmup_prefix (bool, optional): If True, after warmup annealing starts from initial LR. Defaults to False.  
-        t_in_epochs (bool, optional): If set to False, returned lr are None. Defaults to True.
+        t_in_epochs (bool, optional): If set to False, returned lr are None. Defaults to True.  
-        noise_range_t (Union[int, float, List[int, float]], optional): Epoch when noise starts.\
+        noise_range_t (Union[int, float, List[int | float]], optional): Epoch when noise starts.\
-            If list or tuple - epoch range, when noise applied. Defaults to None.
+            If list or tuple - epoch range, when noise applied. Defaults to None.  
-        noise_pct (float, optional): Percentage of noise to add. Defaults to 0.67.
+        noise_pct (float, optional): Percentage of noise to add. Defaults to 0.67.  
-        noise_std (float, optional): Noise standard deviation. Defaults to 1.0.
+        noise_std (float, optional): Noise standard deviation. Defaults to 1.0.  
-        noise_seed (int, optional): Seed to use to add random noise. Defaults to 42.
+        noise_seed (int, optional): Seed to use to add random noise. Defaults to 42.  
-        k_decay (float, optional): Power for k_decay. Defaults to 1.0.
+        k_decay (float, optional): Power for k_decay. Defaults to 1.0.  
-        initialize (bool, optional): Add initial_{field_name} to optimizer param group. Defaults to True.
+        initialize (bool, optional): Add initial_{field_name} to optimizer param group. Defaults to True.  
     """
 
     def __init__(self,
@@ -56,7 +56,7 @@ class CosineLRScheduler(Scheduler):
                  warmup_lr_init: float = 0,
                  warmup_prefix: bool = False,
                  t_in_epochs: bool = True,
-                 noise_range_t: Union[int, float, List[int, float]] = None,
+                 noise_range_t: Union[int, float, List[Union[int, float]]] = None,
                  noise_pct: float = 0.67,
                  noise_std: float = 1.0,
                  noise_seed: int = 42,