diff --git a/unsloth/import_fixes.py b/unsloth/import_fixes.py
index f0dde256c1..bb6996a3e3 100644
--- a/unsloth/import_fixes.py
+++ b/unsloth/import_fixes.py
@@ -101,7 +101,9 @@ if os.environ.get("UNSLOTH_ENABLE_LOGGING", "0") != "1":
     # Skipping import of cpp extensions due to incompatible torch version 2.9.0+cu128 for torchao version 0.15.0
     logging.getLogger("torchao").setLevel(logging.ERROR)
     # SyntaxWarning: invalid escape sequence '\.'
-    warnings.filterwarnings("ignore", message = "invalid escape sequence", category = SyntaxWarning)
+    warnings.filterwarnings(
+        "ignore", message = "invalid escape sequence", category = SyntaxWarning
+    )
 
 
 # Fix up AttributeError: 'MessageFactory' object has no attribute 'GetPrototype'
@@ -549,5 +551,8 @@ def fix_diffusers_warnings():
 def fix_huggingface_hub():
     # huggingface_hub.is_offline_mode got removed, so add it back
     import huggingface_hub
+
     if not hasattr(huggingface_hub, "is_offline_mode"):
-        huggingface_hub.is_offline_mode = lambda: huggingface_hub.constants.HF_HUB_OFFLINE
+        huggingface_hub.is_offline_mode = (
+            lambda: huggingface_hub.constants.HF_HUB_OFFLINE
+        )
diff --git a/unsloth/models/_utils.py b/unsloth/models/_utils.py
index c1f626ec66..1cead3afaf 100644
--- a/unsloth/models/_utils.py
+++ b/unsloth/models/_utils.py
@@ -2386,6 +2386,7 @@ def make_fast_generate_wrapper(original_generate):
     Creates a wrapper around model.generate that checks for incorrect
     vLLM-style usage when fast_inference=False.
     """
+
     @functools.wraps(original_generate)
     def _fast_generate_wrapper(*args, **kwargs):
         # Check for vLLM-specific arguments
@@ -2419,9 +2420,9 @@ def make_fast_generate_wrapper(original_generate):
                     "when `fast_inference=True` (vLLM). Since `fast_inference=False`, you must "
                     "tokenize the input first:\n\n"
                     "  messages = tokenizer.apply_chat_template(\n"
-                    "      [{\"role\": \"user\", \"content\": \"Your prompt here\"}],\n"
+                    '      [{"role": "user", "content": "Your prompt here"}],\n'
                     "      tokenize=True, add_generation_prompt=True,\n"
-                    "      return_tensors=\"pt\", return_dict=True\n"
+                    '      return_tensors="pt", return_dict=True\n'
                     "  )\n"
                     "  output = model.fast_generate(\n"
                     "      **messages.to('cuda'),\n"