| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2026-09-02 | TD(lambda), and a big rework | tslil | |
| NOTE: 6x6 weights are as-yet untrained On the performance side: * switch from linked-list chaining hash table to direct mapping * cache check_win by hash * persist transposition tables across plies (this was a big deal and difficult to puzzle out) * add align directives to the network floats * prepare_nn_input gained a small LUT to avoid branching (i suppose you might call this 1D embedding) * apply mtune=native for the native builds On the correctness side: * Zobrist has now includes the ply parity * the network now predicts something in [-1, 1] using tanh (approximated by polynomials) * the training object is now MSE because we're learning a value estimate * use .clangd instead of compile_commands.json On the training side: * completely forgo human-derived data set * implement td(lambda), self-play, tournaments | |||
| 2023-01-15 | 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 So cnn1986 has been replaced by nn1986, a standard, two-layer, dense nn configured as a binary classifier and (mis)used in that capacity.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. | |||
| 2021-10-22 | --fast-math seems saves some time | tslil clingman | |
| That i can tell at a glance, the largest relevant change this introduces is flush-to-zero for floats. Some testing is required to ensure that the CNN isn't too sensitive to this. Preferable to that option is to simply train it without relying on subnormal floats in the first place. | |||
| 2021-10-05 | 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. | |||
| 2021-02-05 | Renamed binaries | tslil clingman | |
| 2021-02-05 | TEI interface working! | tslil clingman | |
| 2021-01-31 | 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. | |||
| 2021-01-30 | Try to squeeze out a little more performance | tslil 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. | |||
| 2021-01-29 | Purged uninteresting statistics | tslil | |
| 2021-01-29 | Cleaned up build system | tslil clingman | |
| 2021-01-28 | Added license information! | tslil clingman | |
| 2021-01-25 | Stripping debug stuff | tslil clingman | |
| 2021-01-24 | This is the basic idea, there's ≥ 1 bug (generates illegals...) | tslil clingman | |
| 2021-01-24 | Working on action lists to refactor | tslil clingman | |
| 2021-01-24 | Still trying | tslil clingman | |
| | | |||
| 2021-01-23 | Hash collisions | tslil clingman | |
| 2021-01-23 | Attempting Zobrist hashing | tslil clingman | |
| 2021-01-23 | Not correct usage, need to try hashing | tslil clingman | |
| 2021-01-21 | Finally (?) fixed winning avoidance | tslil clingman | |
| 2021-01-20 | Infrastructure more or less set up! | tslil clingman | |
| It remains to tweak ct1986.c and add a .login file to autorun stuff! | |||
| 2021-01-20 | In-lined as much as a i dare | tslil clingman | |
| 2021-01-20 | Closing in on lcd! | tslil clingman | |
| 2021-01-19 | Slight bug | tslil clingman | |
| 2021-01-18 | First steps towards character LCD | tslil clingman | |
| 2021-01-18 | Multiplatform builds! | tslil clingman | |
| 2021-01-18 | Cross-compilation! | tslil clingman | |
| 2021-01-18 | Shed dependency of linenoise | tslil clingman | |
| 2021-01-17 | Revert "Starting work on MCU port" | tslil clingman | |
| This reverts commit 43eec01361eb6174142180c29afef960e041514e. | |||
| 2021-01-17 | It works, but it's too slow | tslil | |
| 2021-01-17 | Starting work on MCU port | tslil clingman | |
| 2021-01-16 | Quick sketch of non-determinism | tslil | |
| 2021-01-15 | Dragon is not separate, fixed concession, allowing playing as b/w | tslil | |
| 2021-01-14 | So after a small tweak, it actually plays well! | tslil | |
| 2021-01-14 | Trying alpha-beta pruning. There's a bug somewhere ... | tslil | |
| 2021-01-13 | Fixed minimax (!), fixed bugs in tak.c | tslil | |
| With minimax of depth 1 the evaluation function seems alright with the current method of training and generating weights | |||
| 2021-01-13 | Trying various things to teach ti | tslil | |
| 2021-01-12 | New weights | tslil | |
| 2021-01-12 | Train on data that is close to the end of the game only | tslil | |
| 2021-01-11 | Close in principle, but there;s a bug in ct1997 &/ pptdb! | tslil | |
| 2021-01-10 | Starting on importing network! | tslil | |
| 2021-01-05 | Gotta go fast | tslil | |
| 2021-01-04 | Fixed indentation and some bugs, stats programme | tslil | |
| 2021-01-03 | Set dialect to C99 | tslil | |
| 2021-01-01 | extract games from playtak.com/games database | tslil | |
| 2020-12-31 | Line-mode mostly complete | tslil | |
| 2020-12-30 | Added linenoise library, uninitialised variable + renaming | tslil | |
| 2020-12-22 | Init | tslil | |
 |
index : ctak | |
| An implementation of Tak and a computer opponent in C | git repository hosting |
| aboutsummaryrefslogtreecommitdiff |