Bence Bagi
Bence Bagi
Thanks!
I'm not sure if this helps with the original problem, but someone might find it useful in the future. You can pass a callable to `GridSearchCV` as the `refit` parameter...
This needs to be more exhaustively tested to make sure it works on real datasets. I have even encountered toy examples where my current implementation didn't work either.
The regularization strength might have to be added to the `L` and `L_max` constants determined. It could also be useful as bound for the convexity when using ridge.
@cjchristopher thanks for the warning! Can you please point to the bugs? I started with a naive implementation, but over time the zoom linesearch ended up being quite similar to...
The tests pass locally with the latest development version of Optax, but will fail with the current release of 0.2.4 which the workflow also uses.
Hi! I ran into the same thing and was able to fix it by updating the docker library to 6.1.2 See the relevant issue and PR here: https://github.com/docker/docker-py/issues/3124
Thank you for the quick response, that sounds fantastic! Absolutely, I will send you an email about the video call.
I will send a PR for this in the afternoon!
Just a reminder that it will need to be added in: - `solvers.__init__` - `solvers._solver_registry` - `conftest.py`