diff --git a/llms/mlx_lm/examples/lora_config.yaml b/llms/mlx_lm/examples/lora_config.yaml
index b616aaf4..1585d69e 100644
--- a/llms/mlx_lm/examples/lora_config.yaml
+++ b/llms/mlx_lm/examples/lora_config.yaml
@@ -59,3 +59,4 @@ lora_parameters:
   rank: 8
   alpha: 16.0
   scale: 10.0
+  dropout: 0.0
diff --git a/llms/mlx_lm/lora.py b/llms/mlx_lm/lora.py
index b89d8f0e..adc426e4 100644
--- a/llms/mlx_lm/lora.py
+++ b/llms/mlx_lm/lora.py
@@ -235,11 +235,6 @@ def run(args, training_callback: TrainingCallback = None):
 
         print(f"Test loss {test_loss:.3f}, Test ppl {test_ppl:.3f}.")
 
-    if args.prompt is not None:
-        raise NotImplementedError(
-            "Please use mlx_lm.generate with trained adapter for generation."
-        )
-
 
 if __name__ == "__main__":
     parser = build_parser()
diff --git a/llms/mlx_lm/tuner/utils.py b/llms/mlx_lm/tuner/utils.py
index f5ce4163..b465146c 100644
--- a/llms/mlx_lm/tuner/utils.py
+++ b/llms/mlx_lm/tuner/utils.py
@@ -32,7 +32,11 @@ def linear_to_lora_layers(
         )
 
     to_lora = lambda lin: LoRALinear.from_linear(
-        lin, r=config["rank"], alpha=config["alpha"], scale=config["scale"]
+        lin,
+        r=config["rank"],
+        alpha=config["alpha"],
+        scale=config["scale"],
+        dropout=config["dropout"],
     )
 
     keys = config.get("keys", None)