yeah. Computers and processors are developed rapidly. Moreover there are some amazing parallelization technologies that can essentially alleviate computational requirements.
I think in a few years training of neural network will barely be a big problem.
You are viewing a single comment's thread from: