Katsuki Ohto
Katsuki Ohto
Do we need ModelWrapper in handyrl/train.py? Moreover, do we need ModelWrapper?
Essentially, variables that can be decided by the worker want to be decided by the worker. Furthermore, if there is a default value on the trainer's side, we would like...
`eval_rate` was not practically useful. `eval_coef` will be better, and it can be hidden config argument.
If results for evaluation episodes that were not completed by the end of the epoch are truncated, not only will the number of results displayed be reduced, but there will...
My sincere thanks for a great library. In recent Shogi AI tournaments, the agents are often sent several moves before they play each other in order to adjust the time...