image_input = ct.TensorType(name="input", shape=(1, 3, ct.RangeDim(default=720), ct.RangeDim(default=720)))
		mlmodel = ct.convert(trace, inputs=[image_input], source='pytorch', convert_to='nn_proto')
GPU support, ANE not working
		image_input = ct.TensorType(name="input", shape=(1, 3, ct.RangeDim(), ct.RangeDim()))
		mlmodel = ct.convert(trace, inputs=[image_input], source='pytorch', convert_to='nn_proto')
CPU ONLY
How to make ANE working with flexible input shape?