style-transfer icon indicating copy to clipboard operation
style-transfer copied to clipboard

Add description for OpenCL, works well as replacement for CUDA

Open naibaf7 opened this issue 8 years ago • 2 comments

I just tested this and the scripts are also fully working with OpenCL, and surprisingly not much slower; i.e. 22 minutes with VGG19 on a W9100/R9 290X/R9 390X/RX 480 and 12 minutes on a GTX1080. Source: https://github.com/BVLC/caffe/tree/opencl Disclaimer: I'm the author of OpenCL Caffe, shameless advertisement to gain more momentum ;)

naibaf7 avatar Dec 05 '16 01:12 naibaf7

Will this work on Windows with caffe and CUDAnn or with OpenCL? I have a 1070 and looking to use it, but not sure why the gpu flag is failing when anything other than 0.

nikAleksandr avatar Dec 20 '16 23:12 nikAleksandr

@nikAleksandr Not sure what you have configured, but yes if configured correctly you can use cuDNN+CUDA on your 1070 with normal Caffe, or you can use OpenCL on either the 1070, or your CPU or onboard GPU, if you have. And it should all work with Windows, if you have Anaconda Python 2.7 and VC++2015 installed. But if something specific does not work for you, I'd need more details to see why.

naibaf7 avatar Dec 21 '16 01:12 naibaf7