diff --git a/examples/MAML-RL/func_maml.py b/examples/MAML-RL/func_maml.py
index ce3d9ef3..6413cc71 100644
--- a/examples/MAML-RL/func_maml.py
+++ b/examples/MAML-RL/func_maml.py
@@ -189,7 +189,7 @@ def main(args):
 
 if __name__ == '__main__':
     parser = argparse.ArgumentParser(
-        description='Reinforcement learning with ' 'Model-Agnostic Meta-Learning (MAML) - Train'
+        description='Reinforcement learning with Model-Agnostic Meta-Learning (MAML) - Train'
     )
     parser.add_argument('--seed', type=int, default=1, help='random seed (default: 1)')
     args = parser.parse_args()
diff --git a/torchopt/_src/optimizer/func/base.py b/torchopt/_src/optimizer/func/base.py
index b2c0ee90..9a4abb05 100644
--- a/torchopt/_src/optimizer/func/base.py
+++ b/torchopt/_src/optimizer/func/base.py
@@ -24,8 +24,8 @@
 
 # mypy: ignore-errors
 class FuncOptimizer:  # pylint: disable=too-few-public-methods
-    """A wrapper class to hold the functional optimizer.
-    It makes it easier to maintain the optimizer states.
+    """A wrapper class to hold the functional optimizer. It makes it easier to maintain the
+    optimizer states.
 
     See Also:
         - The functional Adam optimizer: :func:`torchopt.adam`.
@@ -40,7 +40,7 @@ def __init__(self, impl: GradientTransformation, *, inplace: bool = False) -> No
         Args:
             impl (GradientTransformation): A low level optimizer function, it could be a optimizer
                 function provided by `alias.py` or a customized `chain` provided by `combine.py`.
-            inplace: (default: :data:`False`)
+            inplace (optional): (default: :data:`False`)
                 The default value of ``inplace`` for each optimization update.
         """
         self.impl = impl
@@ -61,6 +61,9 @@ def step(
                 loss that is used to compute the gradients to network parameters.
             params: (tree of torch.Tensor)
                 An tree of :class:`torch.Tensor`\s. Specifies what tensors should be optimized.
+            inplace (optional): (default: :data:`None`)
+                Wether to update the parameters in-place. If :data:`None`, use the default value
+                specified in the constructor.
         """
         if self.optim_state is None:
             self.optim_state = self.impl.init(params)