fix mobilebert_MRPC_qdq (intel#1057)

chensuyue · Jul 21, 2022 · ed5a997 · ed5a997
1 parent cf72656
commit ed5a997
Showing 1 changed file with 3 additions and 1 deletion.
diff --git a/examples/onnxrt/language_translation/mobilebert/quantization/ptq/main.py b/examples/onnxrt/language_translation/mobilebert/quantization/ptq/main.py
@@ -101,7 +101,9 @@
         model = model_optimizer.model
 
         from neural_compressor.experimental import Quantization, common
+        from neural_compressor import options
+        options.onnxrt.qdq_setting.OpTypesToExcludeOutputQuantizatioin = ['MatMul']
         quantize = Quantization(args.config)
         quantize.model = common.Model(model)
         q_model = quantize()
-        q_model.save(args.output_model)
+        q_model.save(args.output_model)