ctak - An implementation of Tak and a computer opponent in C

Age	Commit message (Collapse)	Author
3 days	new neural network arch (faster + better) & minor changes + fixes	tslil clingman
	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 So cnn1986 has been replaced by nn1986, a standard, two-layer, dense nn configured as a binary classifier and (mis)used in that capacity.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.
3 days	Went for more standard unicode symbols	tslil clingman

3 days	Corrected win screen in geminict	tslil clingman

3 days	Enforce quoted ptn string for geminict	tslil clingman

3 days	Welcome geminict!	tslil clingman
	This is a special interface to negamax_cnn1986 which is designed to generate output for use in a CGI tak interface to be used over gemini. Also in this commit is a reformating of the various source files to use the traditional tab width of 8 spaces.
3 days	Corrected generation of training data for 6s	tslil clingman

3 days	Don't generate header for training data + tweaks	tslil
	For some reason it would seem that moving flats to a lower value and increasing the proximity between caps and top flats improves acquisition. Still not great, but every bit counts.
3 days	Change the training data generation a little	tslil clingman
	Although it pains me to say it, ``label smoothing'' appears to be actually work. I'm also currently experimenting with training simply against _all_ games, instead of only bot matches. Once the training finishes i'll pit cttei against itself with old and new weights, hopefully there'll be a noticeable improvement.
3 days	Rename ct_k -> ct, IANAL but ...	tslil

3 days	Small typo in generated output for weights	tslil clingman

3 days	Correct line clearing behaviour, EXIT_FAILURE <~ -1	tslil clingman

3 days	Renamed binaries	tslil clingman

3 days	TEI interface working!	tslil clingman

3 days	Tried some naive iterative deepening. Work on TEI interface next	tslil clingman
	If TEI is implemented, then i could make use of Morten's racetrack (https://github.com/MortenLohne/racetrack) and develop a quantitative measure of the bot's performance. This is the current priority.
3 days	Just some #weightgoals ;)	tslil clingman
	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.
3 days	Fairly important bug fixes to lcdlib, LCD now echoes input!	tslil clingman
	Input polling without line-buffering is done using ncurses, so the buildroot configuration had to change accordingly to include that library. The Makefile changed to accommodate stand-alone building of ct1986 and to include -lcurses where appropriate. There were also some typos about copying ct1986 and ctaklm to the correct directories.
3 days	Purged uninteresting statistics	tslil

3 days	Syntax errors, small tweak to training data generation	tslil clingman

3 days	Added license information!	tslil clingman

3 days	Print progress before recursion rather than after	tslil clingman

3 days	Small oversight	tslil clingman

3 days	Small changes to build ct1986	tslil clingman

3 days	Stable negamax-alpha-beta fail-soft	tslil clingman

3 days	Actually it seems before i was mis-counting	tslil clingman

3 days	There is still a bug, it doesn't appear to be checking enough	tslil clingman

3 days	Stripping debug stuff	tslil clingman

3 days	Still bugs...	tslil clingman

3 days	Still some bugs, standing stone becomes flat at depth4 self-play??	tslil clingman

3 days	Incidental bugfix	tslil clingman

3 days	Memory leak fix	tslil clingman

3 days	This is the basic idea, there's ≥ 1 bug (generates illegals...)	tslil clingman

3 days	Working on action lists to refactor	tslil clingman

3 days	Still trying	tslil clingman
	\|
3 days	Hash collisions	tslil clingman

3 days	Fixed some important bugs	tslil clingman

3 days	Attempting Zobrist hashing	tslil clingman

3 days	no speed increase ...	tslil clingman

3 days	Trying caching to speed things up	tslil clingman

3 days	Should now play out to the end	tslil clingman

3 days	Starting work on LCD library	tslil clingman

3 days	Slight fix to game loop	tslil clingman

3 days	Renaming	tslil clingman

3 days	Many fixes, i think this is actually correct	tslil clingman

3 days	Experimenting	tslil clingman

3 days	Vertical line is just incremented %2, i'll have to manually buffer	tslil clingman
	charlcd.c:621 case '\n': /* * flush the remainder of the current line and * go to the beginning of the next line */ for (; priv->addr.x < lcd->bwidth; priv->addr.x++) lcd->ops->write_data(lcd, ' '); priv->addr.x = 0; priv->addr.y = (priv->addr.y + 1) % lcd->height; charlcd_gotoxy(lcd); break;
3 days	Working on better output + build system refinement	tslil clingman

3 days	Forgot some file arguments	tslil clingman

3 days	Can't stop half-way through turn	tslil clingman

3 days	In-lined as much as a i dare	tslil clingman

3 days	Closing in on lcd!	tslil clingman