diff --git a/llms/mlx_lm/tuner/dpo_trainer.py b/llms/mlx_lm/tuner/dpo_trainer.py
index 4ddc3d2e..ed955e01 100644
--- a/llms/mlx_lm/tuner/dpo_trainer.py
+++ b/llms/mlx_lm/tuner/dpo_trainer.py
@@ -128,7 +128,6 @@ def dpo_loss(
         'chosen_logits_mean': mx.mean(policy_chosen_score)
     }
 
-
     return mx.mean(losses), reward, num_tokens, metrics
 
 
@@ -180,7 +179,6 @@ def evaluate_dpo(
     model,
     reference_model,
     dataset,
-    tokenizer,
     batch_size,
     num_batches,
     beta: float,
@@ -328,7 +326,6 @@ def train_dpo(
                 model=model,
                 reference_model=reference_model,
                 dataset=val_dataset,
-                tokenizer=tokenizer,
                 batch_size=args.batch_size,
                 num_batches=args.val_batches,
                 max_seq_length=args.max_seq_length,