fix

python/llm/src/ipex_llm/transformers/npu_model.py

@@ -411,17 +411,18 @@ def load_low_bit(cls, pretrained_model_name_or_path: str, *model_args, **kwargs)
         if enable_cpp_backend:
             from .npu_models.npu_llm_cpp import load_model_from_file
             from .npu_models.convert import generate
-            import json
             dummy_model = torch.nn.Module()
             try:
                 model_ptr = load_model_from_file(pretrained_model_name_or_path)
+                dummy_model.config = PretrainedConfig.from_dict(config_dict)
                 dummy_model.model_ptr = model_ptr
                 dummy_model.save_directory = pretrained_model_name_or_path
                 dummy_model.kv_len = config_dict['kv_len']
                 dummy_model.vocab_size = config_dict['vocab_size']
             except:
                 invalidInputError(False,
                                   "False to InitLLMPipeline.")
+            dummy_model.eval()
             # patch generate function
             import types
             dummy_model.generate = types.MethodType(generate, dummy_model)