diff options
| author | tslil clingman <tslil@posteo.de> | 2021-01-31 23:33:39 -0500 |
|---|---|---|
| committer | tslil <tslil@posteo.de> | 2026-08-28 19:37:41 +0100 |
| commit | 328c8d1e3094a942d6a2edd933c9cc4ab09daab1 (patch) | |
| tree | 21332661fb48c548b04d03d9a4131abe2070699a /resources | |
| parent | 399d90f1b94717aa0ca5abb1dc43bdb6f4160d13 (diff) | |
Just some #weightgoals ;)
It turns out that while i was training on a 0/1 classification
problem, i was using 2*eval - 1. Training using this function instead,
and on bot-dominated game choices (chosen_player in extract.sh) seems
to have given a better evaluation function. At the least, Morten's
swindle doesn't work anymore.
Diffstat (limited to 'resources')
| -rwxr-xr-x | resources/extract.sh | 6 |
1 files changed, 2 insertions, 4 deletions
diff --git a/resources/extract.sh b/resources/extract.sh index 7fdda28..0e78505 100755 --- a/resources/extract.sh +++ b/resources/extract.sh @@ -2,9 +2,7 @@ db_file=games_anon.db -chosen_players="fwwwwibib archvenison NohatCoder nqeron ManaT -SultanPepper orfane rabbitboy84 applemonkeyman Reid Tayacan Breeze -AaaarghBot" +chosen_players="AaaarghBot Tiltak_Bot" query() { query="(size == $1) and (result != '1-0') and (result != '0-1') and (result != '0-0') and (result != '1/2-1/2')" @@ -62,7 +60,7 @@ fi make pptdb -for size in 5 6; do +for size in 5; do echo extract $size notation,result process $size 400000 training |