Reputation: 4357
I was wondering if there is a way to use my GPU to speed up the training of a network in PyBrain.
Upvotes: 1
Views: 2102
Reputation: 8897
Checkout Pylearn2: https://github.com/lisa-lab/pylearn2
It's a newer library and can run on GPU's with CUDA via cuda-convnet.
Upvotes: 1
Reputation: 169269
Unless PyBrain is designed for that, you probably can't.
You might want to try running your trainer under PyPy if you aren't already -- it's significantly faster than CPython for some workloads. Perhaps this is one of those workloads. :)
Upvotes: 2