Re:Deep Learning and Free Software
by blt on Thursday July 26, 2018 @04:39PM
(#17620834)
It's definitely feasible and widely used for reinforcement learning on low dimensional systems, where the neural networks are small and the simulator is more expensive than backprop. On other hand, deep Q learning from pixels on Atari is practically impossible without GPUs.
[ Parent
| on HN ]
Re:Deep Learning and Free Software
by tree_of_item on Thursday July 26, 2018 @08:05PM
(#17622442)
Huh? I just showed you reinforcement learning from pixels on Atari with CPU...