Xtra-Computing / thundersvm

ThunderSVM: A Fast SVM Library on GPUs and CPUs
Apache License 2.0
1.57k stars 217 forks source link

SVC sv indices #138

Closed berryweinst closed 5 years ago

berryweinst commented 5 years ago

Hi, In Sklearn SVC we had access to the sv indices (model.support_). How can one extract them from the Thunder model?

Thanks.

QinbinLi commented 5 years ago

Hi @berryweinst

We have added the attribute support_. You can update the code and try again. Thanks.

berryweinst commented 5 years ago

Thank you so much for the quick response and the enhancement. Can I ask to update the wheel file as well please? thundersvm-cu90-0.2.0-py3-none-linux_x86_64.whl

jiahuanluo commented 5 years ago

Hi @berryweinst , I have uploaded a new version wheel file to PyPI. You can use pip install thundersvm for Linux with Cuda 9.0 and Python3. Unable to update the wheel on our server so quickly.

Cheers, Locke

berryweinst commented 5 years ago

Thank you so much. Seem to work very well.