Issues while saving a TorchScript Model(Anynet) using torch.jit.save()

[nakul3112]

@lukemelas Hi Luke,

I am a student working on a project on disparity estimation. For that I have trained AnyNet model, however, while converting it to torchscript, I come across some errors. Would you let me know of some reference guides, to annotate the functions or something to be done to custom layers of the network, so that it doesn't give error while saving.

Error format is like this: Could not export Python function call <functionName>. Remove calls to Python functions before export. Did you forget to add @script or @script_method annotation? If this is nn.ModuleList, add it to __constants__.

My Pytorch has following version. Pytorch: 1.9

Any help would be greatly appreciated.

Regards, Nakul