CUDNN support for Convolution Layers

arrayfire / arrayfire-ml

ArrayFire's Machine Learning Library.

BSD 3-Clause "New" or "Revised" License

102 stars 23 forks source link

CUDNN support for Convolution Layers #1

Closed pavanky closed 7 years ago

pavanky commented 9 years ago

Wraps perceptrons, convolutional and softmax layers

Other

[ ] Integrate with cuDNN when possible
[ ] Add a fallback version using cuDNN is not available.
[ ] Investigate alternatives and analyze performance

futurely commented 9 years ago

facebook/fbcunn exploits cuDNN through a limited number of DNN specific operations implemented by the wrapper cudnn.torch. This is probably the least invasive method to integrate with cuDNN.

The fallback CUDA backend of torch is cunn.

UniqueFool commented 8 years ago

FWIW, this may be worth looking at (Mozilla Public License 2.0): https://github.com/hughperkins/DeepCL More specfically: https://github.com/hughperkins/DeepCL/tree/master/cl

pavanky commented 7 years ago

This is going to be solved upstream: https://github.com/arrayfire/arrayfire/issues/1868