Open wenkesj opened 7 years ago
Good idea. Problem is issue #5. We have to solve this first.
I still think I proof of concept would be interesting. I might try that out. It would be a ridiculously valuable resource to be training relatively inexpensively
Amazing work! Sorry for posting in here, I am wondering if this is related to this article: http://www.cnx-software.com/2017/04/30/gpu-accelerated-object-recognition-on-raspberry-pi-3-raspberry-pi-zero/ and if so, whether an example will be made available - it would greatly benefit the open source community and those who cannot afford Nvidia platform - needless to say the feedback from that innovation creates even more innovation.
Any update about running TensorFlow ops on Raspberry Pi GPU?
Apparently not.
Do you think this would be reasonable to implement with Tensorflow as a py_func op? I think it's reasonable to assume it would be worth creating a module as a substitute for sgemm or other heavy ops.