diff --git a/src/rxn/onmt_utils/train_command.py b/src/rxn/onmt_utils/train_command.py
index b316040..1511ba7 100644
--- a/src/rxn/onmt_utils/train_command.py
+++ b/src/rxn/onmt_utils/train_command.py
@@ -1,8 +1,14 @@
+import logging
 from enum import Flag
-from typing import Any, List, Tuple
+from typing import Any, List, Optional, Tuple
 
 from rxn.utilities.files import PathLike
 
+from .model_introspection import get_model_rnn_size
+
+logger = logging.getLogger(__name__)
+logger.addHandler(logging.NullHandler())
+
 
 class RxnCommand(Flag):
     """
@@ -246,7 +252,6 @@ def finetune(
         data: PathLike,
         dropout: float,
         learning_rate: float,
-        rnn_size: int,
         save_model: PathLike,
         seed: int,
         train_from: PathLike,
@@ -256,7 +261,15 @@ def finetune(
         data_weights: Tuple[int, ...],
         report_every: int,
         save_checkpoint_steps: int,
+        rnn_size: Optional[int] = None,
     ) -> "OnmtTrainCommand":
+        if rnn_size is None:
+            # In principle, the rnn_size should not be needed for finetuning. However,
+            # when resetting the decay algorithm for the learning rate, this value
+            # is necessary - and does not get it from the model checkpoint (OpenNMT bug).
+            rnn_size = get_model_rnn_size(train_from)
+            logger.info(f"Loaded the value of rnn_size from the model: {rnn_size}.")
+
         return cls(
             command_type=RxnCommand.F,
             no_gpu=no_gpu,