Do all the things like ++ or -- rants, post your own rants, comment on others' rants and build your customized dev avatarSign Up
-ANGRY-CLIENT-8541100dSame here. I will probably buy a computer or build one myself
-ANGRY-CLIENT-8541100dI've tried to do machine learning e.g. on my laptop.
It just takes way too much resources.
magicMirror4891100dwhy would you do that to yourself?
spin up an amazon machine when you need those.
ilikeglue1569100dI really wish to know why tensor flow doesn't work with opencl on amd cards...
Alice68295100dWhen you want to install something and waste time without knowing what your machine supports instead of looking for it right away.
BigETI50100dI don't believe a long lasting future in proprietary technology like CUDA. Open software designed for multi-purpose computation like OpenCL is the way to go. Also it can run code either on CPU, GPU or even on FPGA.
paranoidAndroid445100dUse DA CLOUD. Google even offers tensor machines. That's their business model fir tensor flow.