Project implementing some optimization routines in Python.
Ø compiled code. Ø bullshit. State of the art performance.
Let me explain this better:
- Ø compiled code, because we don't need to. Numba generates machine-efficient code on the fly. This way, the code remains pure Python, which makes it both easier to develop and to deploy.
- Ø bullshit. Here you will only find methods that work out of the box. No more tweaking hyperparameters to make the method converge. Pinkie swear.
- State of the art performance. We are constantly working on making the methods as fast as possible.