Problem of implementation

[w-shuw]

Hi, thanks for your great work!
There are something wrong when I deploy model with C++. When i use torch.jit.trace to export the model, the problem was arised like below:
Could not export Python function call 'xxx'. Remove calls to Python functions before export. Did you forget to add @script or @script_method annotation? If this is a nn.ModuleList, add it to __constants__:
If you learn about it, could you help me with it.