Corrected typo in return_full_text parameter

Changed eturn_full_text=False to return_full_text=False on line 80 to fix a typo in the parameter name.

README.md

@@ -77,7 +77,7 @@ pipeline = transformers.pipeline(
     model=model_id,
     max_new_tokens=600,
     device="cuda",
-    eturn_full_text=False
+    return_full_text=False
 )
 
 messages = [