microsoftml.gpu_math: Acceleration with NVidia CUDA
Usage
microsoftml.gpu_math(gpu_id: numbers.Real = -1,
cu_dnn: bool = False, cu_dnn_algo: str = 'ImplicitPrecompGemm')
Description
NVidia CUDA implementation.
Arguments
gpu_id
GPU device id (settings).
cu_dnn
Use cuDNN on GPU (settings).
cu_dnn_algo
cuDNN optimization options (settings).
See also
Feedback
Coming soon: Throughout 2024 we will be phasing out GitHub Issues as the feedback mechanism for content and replacing it with a new feedback system. For more information see: https://aka.ms/ContentUserFeedback.
Submit and view feedback for