I have a project that must use Libtorch to train a model for one portion.
If I could use PyTorch for the other portion it would make things easier.
If I have a model saved in the C++ version of torch can I get the model loaded in Python’s version (Pytorch, of course) and can I do the opposite? Load a model from Pytorch into the C++ version Libtorch?
I know I should try, and find this out for myself. It’s taking me a while to get to know Libtorch, so maybe someone can get me a faster answer here. Maybe the models are usable one way but not the other. Maybe not at all. Maybe yes to all.