caffe2 icon indicating copy to clipboard operation
caffe2 copied to clipboard

Published 20 hours ago • verified publisher icon facebookarchive

OpenGL implementation

Open weiyichang opened this issue 7 years ago • 2 comments

Hi, Recently, I notice that the OpenGL backend (for mobile) has been released, and I also discover that the main difference is using GLPredictor. But the input and output of GLPredictor are different from the Predictor, and the usage of GLImage/GLImageVector is lack of information. I am not sure how to setup the input image data (e.g., YUV buffer in AICamera example) for GLImage/GLImageVector.

So is there any example with OpenGL backend, or any suggestion about the transformation between TensorCPU and GLImage? Thanks!!!

weiyichang avatar Sep 12 '17 06:09 weiyichang

Moreover, does OpenGL backend supports FC layers and Element-wise layer ?

weiyichang avatar Sep 14 '17 09:09 weiyichang

I wonder that too. Caffe2 have a lot of publicity about creating a real-time application with the opengl backend, but nobody knows how to do it.

cheetah132 avatar Dec 02 '17 16:12 cheetah132