onnx: Got error with upsampling operation

my implementation is below…

import onnx
from onnx import helper
from onnx import TensorProto

import numpy as np

data = np.array([[[
    [1, 2],
    [3, 4],
]]], dtype=np.float32)

scales = np.array([1.0, 1.0, 2.0, 3.0], dtype=np.float32)

inputs = list()
input_tensor = helper.make_tensor_value_info(
    'x', TensorProto.FLOAT, data.shape)
scale_tensor = helper.make_tensor_value_info(
    'scales', TensorProto.FLOAT, (4,))
inputs.append(input_tensor)
inputs.append(scale_tensor)

initializers = list()
initializers.append(helper.make_tensor('scales', TensorProto.FLOAT, scales.shape, scales))

graph = helper.make_graph(
            nodes=[helper.make_node("Upsample", ['x', 'scales'], ['y'])],
            name='UPSAMPLE-TEST',
            inputs=inputs,
            outputs=[],
            initializer=initializers)

model_def = helper.make_model(graph)
onnx.checker.check_model(model_def)

and I got the output

onnx.onnx_cpp2py_export.checker.ValidationError: Input size 2 not in range [min=1, max=1].

==> Context: Bad node spec: input: "x" input: "scales" output: "y" op_type: "Upsample"

I can’t understand error message…any hint? Thanks!

my onnx version is 1.3.0 with python 2.7.15.