0

My laptop is Macbook Pro 2016 with graphic card Radeon Pro 460. Is there anyway that I can leverage my GPU to speed up tensorflow?

I understand that CUDA is for NVIDIA card, and CUDA version of TF doesn't work. Is there any other tool that I can use GPU to run TF such as OpenCL?

Update: One potential answer is that: "You can run Tensorflow on a Macbook Pro 2016 using tf-coriander . Disclosure: I'm the author. – Hugh Perkins 2 days ago"

drbombe
  • 609
  • 7
  • 15
  • 5
    You can run Tensorflow on a Macbook Pro 2016 using [tf-coriander](https://github.com/hughperkins/tf-coriander) . Disclosure: I'm the author. – Hugh Perkins Jun 09 '17 at 02:10
  • GREAT! can you put it as an answer, then other people can easily to see it – drbombe Jun 09 '17 at 02:13

1 Answers1

3

Presumably, you've read: https://www.tensorflow.org/install/install_mac

You cannot "trick your computer" and the suggestion that you might is perhaps contributing to your downvotes.

If you want to learn more about the potential to run TF on your current hardware have a look at the in-progress work being done on TF OpenCL (which supports AMD cards): https://www.google.com/search?q=tensorflow%20for%20opencl&rct=j that said, I think the OpenCL work for linux is much further along than for macos. Good luck!

ruoho ruotsi
  • 1,283
  • 14
  • 13