E:\python_test\shufflenetv2-onnx\ShuffleNetV2.py:111: TracerWarning: Converting a tensor to a Python index might cause the trace to be incorrect. We can’t record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
x1 = x[:, :(x.shape[1] // 2), :, :]
E:\python_test\shufflenetv2-onnx\ShuffleNetV2.py:112: TracerWarning: Converting a tensor to a Python index might cause the trace to be incorrect. We can’t record the data flow of Python values, so this value will be treated as a constant in the future. This means that the trace might not generalize to other inputs!
x2 = x[:, (x.shape[1] // 2):, :, :]
%186 : Float(1!, 24, 28, 28) = onnx::Sliceaxes=[1], ends=[24], starts=[0], scope: ShuffleNetV2/Sequential[features]/InvertedResidual[1]
%187 : Float(1!, 24, 28, 28) = onnx::Sliceaxes=[2], ends=[9223372036854775807], starts=[0], scope: ShuffleNetV2/Sequential[features]/InvertedResidual[1]
%188 : Float(1!, 24, 28, 28) = onnx::Sliceaxes=[3], ends=[9223372036854775807], starts=[0], scope: ShuffleNetV2/Sequential[features]/InvertedResidual[1]
%189 : Float(1, 48, 28, 28) = onnx::Sliceaxes=[0], ends=[9223372036854775807], starts=[0], scope: ShuffleNetV2/Sequential[features]/InvertedResidual[1]
%190 : Float(1!, 24, 28, 28) = onnx::Sliceaxes=[1], ends=[9223372036854775807], starts=[24], scope: ShuffleNetV2/Sequential[features]/InvertedResidual[1]
%191 : Float(1!, 24, 28, 28) = onnx::Sliceaxes=[2], ends=[9223372036854775807], starts=[0], scope: ShuffleNetV2/Sequential[features]/InvertedResidual[1]
%192 : Float(1!, 24, 28, 28) = onnx::Sliceaxes=[3], ends=[9223372036854775807], starts=[0], scope: ShuffleNetV2/Sequential[features]/InvertedResidual[1]
When i transform a pytorch model to onnx, i got this issue. At the same time, the Op Slice was transform error, so how can i fix this issue?