diff options
| author | tslil clingman <tslil@posteo.de> | 2023-01-15 16:03:37 +0100 |
|---|---|---|
| committer | tslil <tslil@posteo.de> | 2026-08-28 19:37:41 +0100 |
| commit | ee216c008a188a9436fedb85c70ee5d1719733b1 (patch) | |
| tree | f1d8fa5efd71851dd4f8d3e2b26b1f95a6086cb9 /include/negamax.h | |
| parent | 7cf3a656d0c923dd92025c09747461f2f2d1bed0 (diff) | |
new neural network arch (faster + better) & minor changes + fixes
Gone is the convolutional neural network, for it turns out not only is
it more difficult to train, but all of the extra information about
board layers didn't make much of a difference at this size.
So cnn1986 has been replaced by nn1986, a standard, two-layer, dense
nn configured as a binary classifier and (mis)used in that capacity.
Note: total number of parameters is unchanged.
HARK: this new nn exposes a bug somewhere in ctak. Run ctlm with
self-play to see the completely borked board state at the end.
Diffstat (limited to 'include/negamax.h')
| -rw-r--r-- | include/negamax.h | 6 |
1 files changed, 3 insertions, 3 deletions
diff --git a/include/negamax.h b/include/negamax.h index 47f4ede..e491baa 100644 --- a/include/negamax.h +++ b/include/negamax.h @@ -20,7 +20,7 @@ #include <tak.h> #include <actions.h> -#include <cnn1986.h> +#include <nn1986.h> #include <tt_llcht.h> #include <zobrist.h> @@ -33,8 +33,8 @@ extern char negamax_ptn[9]; extern uint8_t negamax_search_depth; extern void negamax_display_progress(const uint8_t cur_depth, - const uint8_t init_depth, - const uint32_t length); + const uint8_t init_depth, + const uint32_t length); // =================================================================== // Methods |