What’s the fastest way to train NNs?

[minwang-ai]

Hi Sylvain,

Since the experiments and blog are from 2018, is that still the fast way to train networks now?

How do you choose the fixed epoch of one circle policy?

Thanks!

Best wishes, Min