search

NeymarL / ChineseChess-AlphaZero

Implement AlphaZero/AlphaGo Zero methods on Chinese chess.
https://cczero.org
GNU General Public License v3.0
1.09k stars 344 forks source link

How to use sl to achieve "faster training"? #44

Open ztqwerty opened 4 years ago

ztqwerty commented 4 years ago

@NeymarL 大神，excellent work, and very nice and clean code! Best Chinese chess repo I've ever seen.

Questions regarding training:

  1. Could you elaborate on how to use the supervised learning worker to accelerate training procedure?
  2. How do you crawl training data from onegreen? Can you share some experience / tips?

Since your server is down (any possibility to be back on again?), I am planning to explore training locally and see if I can get a better model

charleschen2006 commented 2 years ago

It seems that there is no server code here? how can I run a server and train it myself?