Update _process_sequence() docstring (#219)

Small modification in the documentation of `_process_sequence` to reflect the correct order (`hidden_state`, `cell_state`) of `lstm_states`.

See order here: https://pytorch.org/docs/stable/generated/torch.nn.LSTM.html

sb3_contrib/common/recurrent/policies.py

@@ -170,7 +170,7 @@ def _process_sequence(
         Do a forward pass in the LSTM network.
 
         :param features: Input tensor
-        :param lstm_states: previous cell and hidden states of the LSTM
+        :param lstm_states: previous hidden and cell states of the LSTM, respectively
         :param episode_starts: Indicates when a new episode starts,
             in that case, we need to reset LSTM states.
         :param lstm: LSTM object.