How ?
You'd run it on your hardware and choose what data to train it on.
Please Login to reply.
Can just crowdsource training and then everyone uses the resulting model
Maybe we would need much more personal computing power ? Or could it be shared power run on nodes. So the computation is shared?
