update

decoder_model_merged_quantized.onnx.onnx

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:fe9751468ddf78017d5bbb317d1162049b57bc8b11321dcf3756907835245e42
+size 20201510

quantifiy.py

@@ -0,0 +1,11 @@
+import onnx
+from onnxruntime.quantization import quantize_dynamic, QuantType
+
+# Load the original init_decoder_model
+model_path = "decoder_model.onnx"
+quantized_model_path = "decoder_quantized.onnx"
+
+# Perform dynamic quantization on the model
+quantize_dynamic(model_path, quantized_model_path, weight_type=QuantType.QInt8)
+
+print(f"Quantized model saved to {quantized_model_path}")