diff --git a/tests/neural_compressor/test_optimization.py b/tests/neural_compressor/test_optimization.py
index bd7b6519c6..2362e50578 100644
--- a/tests/neural_compressor/test_optimization.py
+++ b/tests/neural_compressor/test_optimization.py
@@ -375,7 +375,7 @@ def calibration_fn(p_model):
                 save_directory=tmp_dir,
                 save_onnx_model=False,
             )
-            model = INCModelForCausalLM.from_pretrained(tmp_dir, export=True)
+            model = INCModelForCausalLM.from_pretrained(tmp_dir)
 
         pre_outputs = quantizer._quantized_model.generate(
             **tokens, do_sample=False, num_beams=1, temperature=0.9, min_length=20, max_length=20