Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

The tree method gpu_hist is deprecated since 2.0.0 #54

Open
Moelf opened this issue Nov 12, 2024 · 2 comments
Open

The tree method gpu_hist is deprecated since 2.0.0 #54

Moelf opened this issue Nov 12, 2024 · 2 comments

Comments

@Moelf
Copy link
Contributor

Moelf commented Nov 12, 2024

The tree method `gpu_hist` is deprecated since 2.0.0. To use GPU training, set the `device` parameter to CUDA instead.

but currently this package doesn't support device="gpu"

@ablaom
Copy link
Member

ablaom commented Nov 13, 2024

Mmm. The pattern elsewhere in MLJ is to have an acceleration hyperparameter, which takes an instance of AbstractResource, provided by ComputationalResources.jl. Examples are CPU1() ("cpu") and CUDALibs() ("gpu"). On the other hand, we generally try to mirror the original names, so should still provide device. We could provided both, but then there's a little logic needed to deal with any ambiguity/conflict.

@Moelf
Copy link
Contributor Author

Moelf commented Nov 21, 2024

I can try a PR but because there's no CI for this I'm a bit hesitant to blindly try it

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants