These 9op are not supported：{Shape,Gather,Unsqueeze,Equal,Cast,TopK,Expand,OneHot,ReduceSum,}

[fyq00]

I know 3op : Shape, Gather, Unsqueeze can be change by simplifier onnx too.
But here are still 6op I need.
My model need to use heatmap prediction. I use function of pytorch :
1 torch.topk() to sort pixel scores.
2 OPtion : (Tensor1 == Tensor2) to zero the other unimportant pixels ....
3 torch.expand() function ~
and so on ~

But it's seem like Tengine still not support this op.

I dont know how to solve this problem~
Can you help me ~

[liqi-c]

topk we support now , but maybe not support in onnx format , Maybe you can try convert you onnx model to tengine models format , and then we can know which op is not support , We will add it later
you alse can reference the tools/plugin/serializer/onnx/onnx_serializer.cpp to check which op we support .