[jnwatson]

Training is what the cloud is for.

[wongarsu]

That makes a $170 board that can also do training look dirt cheap in comparison

[lawrenceyan]

Good luck training anything in any reasonable time on it.

[R0b0t1]

Useful for adapting existing models. Not everything needs millions of hours of input.

[tachyonbeam]

If you want to train yet-another-convnet sure, but there could be applications where you want to train directly on a robot with live data, as in interactive learning.

See this paper for an example of interactive RL: https://arxiv.org/abs/1807.00412

[suyash]

or a highly rigged machine, this looks more for fast real time ML inference on the edge