HCha/Stockfish
1
0
Fork 0
mirror of https://github.com/HChaZZY/Stockfish.git synced 2025-12-23 18:46:59 +08:00
Code Issues Packages Projects Releases Wiki Activity

Add missing docs.

Browse Source
This commit is contained in:
Tomasz Sobczyk
2020-11-22 18:08:14 +01:00
committed by nodchip
parent 9030020a85
commit 45e3335ee8
2 changed files with 6 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
docs/learn.md
View File

@@ -64,6 +64,10 @@ Currently the following options are available:
`newbob_decay` - learning rate will be multiplied by this factor every time a net is rejected (so in other words it controls LR drops). Default: 0.5 (no LR drops)
`assume_quiet` - this is a flag option. When specified learn will not perform qsearch to reach a quiet position.
`smart_fen_skipping` - this is a flag option. When specified some position that are not good candidates for teaching are skipped. This includes positions where the best move is a capture or promotion, and position where a king is in check.
`newbob_num_trials` - determines after how many subsequent rejected nets the training process will be terminated. Default: 4.
`auto_lr_drop` - every time this many positions are processed the learning rate is multiplied by `newbob_decay`. In other words this value specifies for how many positions a single learning rate stage lasts. If 0 then doesn't have any effect. Default: 0.