There was an error while loading. Please reload this page.
8fea425
CUDA has been enabled.
Gradient evals are now faster, for both OpenCL and CUDA.