Skip to content
This repository was archived by the owner on Aug 7, 2023. It is now read-only.

small fixes for cupy11#43

Open
quasiben wants to merge 3 commits intorapidsai:branch-22.12from
quasiben:fix-cupy11
Open

small fixes for cupy11#43
quasiben wants to merge 3 commits intorapidsai:branch-22.12from
quasiben:fix-cupy11

Conversation

@quasiben
Copy link
Member

cc @trivialfis @dantegd

This are fixes needed for CuPy 11 which will also be consumed by DLFW -- i'm using them in 22.10 as well

np.float32, xgb.DeviceQuantileDMatrix, np.nan)
with pytest.raises(xgb.core.XGBoostError):
xgb.train({'tree_method': 'gpu_hist', 'gpu_id': 1},
'fail_on_invalid_gpu_id': True},
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

This should fail without additional parameters assuming there is more than 1 device available during testing.

quasiben and others added 2 commits October 18, 2022 11:18
Co-authored-by: Jiaming Yuan <jm.yuan@outlook.com>
remove unnecessary option
@quasiben
Copy link
Member Author

Do we also need to submit an upstream patch for this as well ?

@trivialfis
Copy link
Member

Do we also need to submit an upstream patch for this as well ?

Thank you for the fix. It's fixed in upstream (but with some refactoring so this PR is still the best option for rapids).

@quasiben
Copy link
Member Author

quasiben commented Nov 2, 2022

@trivialfis / @dantegd can this be merged in ?

@trivialfis
Copy link
Member

I don't have write access

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants