aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
47 hoursCorrect line clearing behaviour, EXIT_FAILURE <~ -1tslil clingman
47 hoursRenamed binariestslil clingman
47 hoursTEI interface working!tslil clingman
47 hoursTried some naive iterative deepening. Work on TEI interface nexttslil 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.
47 hoursJust 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.
47 hoursFairly 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.
47 hoursTry to squeeze out a little more performancetslil clingman
``Common wisdom'' dictates that placements are often better than stack moves, so we bias the generated move list in this fashion. Seems to be a little faster.
47 hoursPurged uninteresting statisticstslil
47 hoursSyntax errors, small tweak to training data generationtslil clingman
47 hoursCleaned up build systemtslil clingman
47 hoursSkeleton readmetslil clingman
47 hoursAdded license information!tslil clingman
47 hoursPrint progress before recursion rather than aftertslil clingman
47 hoursTypotslil clingman
47 hoursMerge branch 'zobrist'tslil clingman
47 hoursSmall oversighttslil clingman
47 hoursRemoved treap in favour of linked-list chained hash tabletslil clingman
47 hoursSmall changes to build ct1986tslil clingman
47 hoursIt would appear that any function call whatsoever is slower :/tslil clingman
For now we'll stay with directly recomputing it at each non-terminal node
47 hoursStoring best moves!tslil clingman
47 hoursI don't have the presence of mind to debug this right nowtslil clingman
47 hoursThis matches alpha-beta!tslil clingman
47 hoursSomehting along these lines, i'm tiredtslil clingman
47 hoursStable negamax-alpha-beta fail-softtslil clingman
47 hoursSmall oversighttslil clingman
47 hoursOnce again, adding TT changes the outcometslil clingman
47 hoursActually it seems before i was mis-countingtslil clingman
47 hoursThere is still a bug, it doesn't appear to be checking enoughtslil clingman
47 hoursStripping debug stufftslil clingman
47 hoursIt was a silly typo! Hoorah!tslil clingman
47 hoursStill bugs...tslil clingman
47 hoursStill some bugs, standing stone becomes flat at depth4 self-play??tslil clingman
47 hoursThere's still something wrongtslil clingman
47 hoursIncidental bugfixtslil clingman
47 hoursThis looks better to me and confirms scribbles on papertslil clingman
47 hoursI don't understand why action_list.c:306 != 259tslil clingman
47 hoursThere are still some bugs in the undo almost surely...tslil clingman
47 hoursLots of bugfixes, mostly uint vs int. Still weirdness in gametslil clingman
47 hoursMemory leak fixtslil clingman
47 hoursThis is the basic idea, there's ≥ 1 bug (generates illegals...)tslil clingman
47 hoursWorking on action lists to refactortslil clingman
47 hoursStill tryingtslil clingman
|
47 hoursHash collisionstslil clingman
47 hoursFixed some important bugstslil clingman
47 hoursAttempting Zobrist hashingtslil clingman
47 hoursNot correct usage, need to try hashingtslil clingman
47 hoursno speed increase ...tslil clingman
47 hoursTrying caching to speed things uptslil clingman
47 hoursThanks compile warnings!tslil clingman
47 hoursShould now play out to the endtslil clingman