do we need train and test the code with same wrapper?

[tiru1930]

i have model which is trained using torch directly now , i have written prediction class in torch which i want to call using python (pytorchhelpers) , but i got below error
PANIC: unprotected error in call to Lua API (/home/tiru/torch/install/share/lua/5.1/torch/File.lua:343: unknown Torch class <torch.CudaTensor>)

[hughperkins]

It looks like it cannot find torch.CudaTensor class/file for some reason. You might need to hack around with the LUA_CPATH and similar.