You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
An autograd engine using tensors/matrices instead of scalar values.
Due to Python's GIL, it is impossible to gain the benefit of using matrix transforms to parallelise the backward pass without creating a c extension.
Mainly for educational purposes