ONNX ConstantOfShape Operator

I have a model that results in some ConstantOfShape operators when exporting to ONNX. This operator isn’t widely supported in a lot of the conversion tools (CoreML, TensorFlow, TensorRT, ect) however it doesn’t appear to be a common problem looking at the lack of issues when Googling.

Creating a tensor of float 0 seems like a pretty common operation to do, what am I missing?