diff --git a/bitsandbytes/optim/adamw.py b/bitsandbytes/optim/adamw.py
index a32394bd5..5f225c9ad 100644
--- a/bitsandbytes/optim/adamw.py
+++ b/bitsandbytes/optim/adamw.py
@@ -26,7 +26,7 @@ def __init__(
         Base AdamW optimizer.
 
         Arguments:
-            params (`torch.tensor`):
+            params (`torch.Tensor`):
                 The input parameters to optimize.
             lr (`float`, defaults to 1e-3):
                 The learning rate.
@@ -87,7 +87,7 @@ def __init__(
         8-bit AdamW optimizer.
 
         Arguments:
-            params (`torch.tensor`):
+            params (`torch.Tensor`):
                 The input parameters to optimize.
             lr (`float`, defaults to 1e-3):
                 The learning rate.
@@ -159,7 +159,7 @@ def __init__(
         32-bit AdamW optimizer.
 
         Arguments:
-            params (`torch.tensor`):
+            params (`torch.Tensor`):
                 The input parameters to optimize.
             lr (`float`, defaults to 1e-3):
                 The learning rate.
@@ -219,7 +219,7 @@ def __init__(
         Paged AdamW optimizer.
 
         Arguments:
-            params (`torch.tensor`):
+            params (`torch.Tensor`):
                 The input parameters to optimize.
             lr (`float`, defaults to 1e-3):
                 The learning rate.
@@ -241,8 +241,6 @@ def __init__(
                 Adapts clipping threshold automatically by tracking the last 100 gradient norms and clipping the gradient at a certain percentile to improve stability.
             block_wise (`bool`, defaults to `True`):
                 Whether to independently quantize each block of tensors to reduce outlier effects and improve stability.
-            is_paged (`bool`, defaults to `False`):
-                Whether the optimizer is a paged optimizer or not.
         """
         super().__init__(
             "adam",
@@ -279,7 +277,7 @@ def __init__(
         Paged 8-bit AdamW optimizer.
 
         Arguments:
-            params (`torch.tensor`):
+            params (`torch.Tensor`):
                 The input parameters to optimize.
             lr (`float`, defaults to 1e-3):
                 The learning rate.
@@ -303,8 +301,6 @@ def __init__(
                 Adapts clipping threshold automatically by tracking the last 100 gradient norms and clipping the gradient at a certain percentile to improve stability.
             block_wise (`bool`, defaults to `True`):
                 Whether to independently quantize each block of tensors to reduce outlier effects and improve stability.
-            is_paged (`bool`, defaults to `False`):
-                Whether the optimizer is a paged optimizer or not.
         """
         # Validate unsupported parameters
         if amsgrad:
@@ -350,7 +346,7 @@ def __init__(
         Paged 32-bit AdamW optimizer.
 
         Arguments:
-            params (`torch.tensor`):
+            params (`torch.Tensor`):
                 The input parameters to optimize.
             lr (`float`, defaults to 1e-3):
                 The learning rate.
@@ -372,8 +368,6 @@ def __init__(
                 Adapts clipping threshold automatically by tracking the last 100 gradient norms and clipping the gradient at a certain percentile to improve stability.
             block_wise (`bool`, defaults to `True`):
                 Whether to independently quantize each block of tensors to reduce outlier effects and improve stability.
-            is_paged (`bool`, defaults to `False`):
-                Whether the optimizer is a paged optimizer or not.
         """
         super().__init__(
             "adam",