Reputation: 71
I have a training dataset of shape (120k+, 14) and I'm trying to use GridSearch for Random Forest with 2500 trees.
Just wondering why my Kaggle notebook never uses the GPU even when CPU usage is over 100%???
Do I have to import something or add some code to start the GPU?
Upvotes: 3
Views: 2315
Reputation: 1042
GPUs are only helpful if you are using code that takes advantage of GPU-accelerated libraries (e.g. TensorFlow, PyTorch, etc.) mostly for training deep learning models. scikit-learn, which you're using here, doesn't support GPUs (and it probably won't). Unless you want to use a different framework, I'd recommend turning the GPU off.
Upvotes: 1