Hi all,
I am converting our model to torchscript using torch.jit.script, but when I execute it, it prints out several lines of warning statement as below.
[W register_special_ops.cpp:229] Warning: Creating a tensor from an empty intlist will create a tensor of default floating point type (currently Float) in python but a tensor of type int in torchscript.
Model’s result is the same with the eager model. I searched on pytorch git repo but couldn’t locate where this warning message is trigged and how to solve or locate the place that the tensor was created from an empty list from our code either. I am using pytorch=12.1 from conda.
Please help, thanks in advance!