Fix LoRA adapter config - clear base_model_name_or_path after applying PEFT

asr_modeling.py

@@ -225,6 +225,10 @@ class ASRModel(PreTrainedModel, GenerationMixin):
                     )
                     model.language_model = get_peft_model(model.language_model, lora_config)
 
+                    # Clear base_model_name_or_path so PEFT doesn't save a reference
+                    # to the base LLM. See _setup_lora for details.
+                    model.language_model.peft_config["default"].base_model_name_or_path = None
+
             return model
         finally:
             cls._is_loading_from_pretrained = False
@@ -393,6 +397,11 @@ class ASRModel(PreTrainedModel, GenerationMixin):
         )
         self.language_model = get_peft_model(self.language_model, lora_config)
 
+        # Clear base_model_name_or_path so PEFT doesn't save a reference to the
+        # base LLM (e.g. Qwen). This prevents pipeline() from redirecting to the
+        # wrong model. The correct path gets set during save_pretrained/push_to_hub.
+        self.language_model.peft_config["default"].base_model_name_or_path = None
+
     def _init_tokenizer(self, config: ASRConfig):
         """Initialize tokenizer with audio token."""
         self.tokenizer = AutoTokenizer.from_pretrained(config.text_model_id, trust_remote_code=True)