This PR fixes an error mentioned in #2.
This was happening due to target tensor not being transferred to GPU.
This PR also introduces an API change to Backprop.calculate_gradient method.
Users can now explicitly set which device to use via use_gpu flag, which is set to False by default.
This PR fixes an error mentioned in #2. This was happening due to
targettensor not being transferred to GPU.This PR also introduces an API change to
Backprop.calculate_gradientmethod. Users can now explicitly set which device to use viause_gpuflag, which is set toFalseby default.