# quantize.py from onnxruntime.quantization import quantize_dynamic, QuantType print("Quantizing bge_m3_onnx/model.onnx to INT8...") quantize_dynamic( model_input="bge_m3_onnx/model.onnx", model_output="bge_m3_onnx/model_quantized.onnx", weight_type=QuantType.QUInt8 ) print("Quantization complete: bge_m3_onnx/model_quantized.onnx created!")