fcde897b8e2f7e9380035e4fa8c53f9d1382d0ca,onnxutils/onnxconverter_common/onnx_ops.py,,apply_upsample,#Any#Any#Any#Any#Any#Any#Any#,532

Before Change

            inputs = [input_name[0], scales_tensor_name]
            op_version = 9

    container.add_node("Upsample", inputs, output_name, op_version=op_version, **attrs)

After Change

    :param mode: nearest or linear
    :param scales: an integer list of scaling-up rate of all input dimensions
    """
    if container.target_opset < 10:
        name = _create_name_or_use_existing_one(scope, "Upsample", operator_name)
        inputs = input_name
        attrs = {"name": name}
        if container.target_opset < 7:
            if len(scales) != 4:
                raise ValueError("Need to specify a 4-element list the the scales of N-, C-, H-, and W-axes")
            attrs["height_scale"] = float(scales[2])
            attrs["width_scale"] = float(scales[3])
            attrs["mode"] = mode.upper()
            op_version = 1
        else:
            attrs["mode"] = mode.lower()
            if container.target_opset < 9:
                attrs["scales"] = list(map(float, scales))
                op_version = 7
            else:
                // scales moved from attribute to input in opset 9
                scales_tensor_name = scope.get_unique_variable_name(name + "_scales")
                container.add_initializer(scales_tensor_name, onnx_proto.TensorProto.FLOAT, [len(scales)], scales)
                inputs = [input_name[0], scales_tensor_name]
                op_version = 9

        container.add_node("Upsample", inputs, output_name, op_version=op_version, **attrs)
    else:
        // TODO, we need verify this after onnx opset 10 release
        name = _create_name_or_use_existing_one(scope, "Resize", operator_name)
        attrs = {"name": name}
        attrs["mode"] = mode.lower()

        scales_tensor_name = scope.get_unique_variable_name(name + "_scales")
        container.add_initializer(scales_tensor_name, onnx_proto.TensorProto.FLOAT, [len(scales)], scales)
        inputs = [input_name[0], scales_tensor_name]
        op_version = 10

        container.add_node("Resize", inputs, output_name, op_version=op_version, **attrs)

Instances