This website requires JavaScript.
107ec1b766
final checkpoint
Christoffer Müller Madsen
2020-01-28 21:44:34 +0100
221e83abd7
Sorry stoffer-boi
Alexander Munch-Hansen
2018-06-15 01:19:51 +0200
a3b4a7480c
Changed it to 1,0 and added a bunch of server stuff
Alexander Munch-Hansen
2018-06-14 23:49:58 +0200
ea4efc5a2b
Updated server code.
Alexander Munch-Hansen
2018-06-07 21:36:06 +0200
63f3e78e9b
fix run_all_tests
Christoffer Müller Madsen
2018-05-26 23:47:10 +0200
02b9e0e0fc
fix 0-ply tests
Christoffer Müller Madsen
2018-05-26 23:36:23 +0200
4ca60f1d4b
fixes to board
Christoffer Müller Madsen
2018-05-24 21:26:40 +0200
445a8435b1
fix network.py
Christoffer Müller Madsen
2018-05-24 17:04:49 +0200
9758ed3fbd
c, no-eager
Christoffer Müller Madsen
2018-05-24 17:04:14 +0200
32a2631ed0
fix bearing off bug
Christoffer Müller Madsen
2018-05-24 16:59:07 +0200
26c0b469eb
restore restore_model
Christoffer Müller Madsen
2018-05-22 20:49:10 +0200
f170bad9b1
tesauro fat and diffs in values
Alexander Munch-Hansen
2018-05-22 15:39:14 +0200
6e061171da
rm TODO
Christoffer Müller Madsen
2018-05-22 15:38:00 +0200
40c228ef01
pubeval tests
Christoffer Müller Madsen
2018-05-22 15:36:23 +0200
c2c6c89e9f
Merge branch 'experimentation' into 'master'
Christoffer Müller Madsen
2018-05-22 13:16:10 +0000
b7708b3675
train-evaluate-save
Christoffer Müller Madsen
2018-05-22 15:15:36 +0200
bad870c27a
update 0-ply-tests
Christoffer Müller Madsen
2018-05-22 15:15:15 +0200
653d6e30a8
add missing comma
Christoffer Müller Madsen
2018-05-22 15:12:47 +0200
7e51b44e33
Merge branch 'experimentation' into 'master'
Christoffer Müller Madsen
2018-05-22 13:12:10 +0000
1fd6c35baa
Merge branch 'master' into 'experimentation'
Christoffer Müller Madsen
2018-05-22 13:11:43 +0000
d426c1c3b5
tesauro fat and diffs in values
Alexander Munch-Hansen
2018-05-22 15:10:41 +0200
5ab144cffc
add git commit status to all logs
Christoffer Müller Madsen
2018-05-22 14:44:13 +0200
cef8e54709
Merge branch 'master' of gitfub.space:Pownie/backgammon
Christoffer Müller Madsen
2018-05-22 14:37:46 +0200
2efbc446f2
log git commit status in evaluation logs
Christoffer Müller Madsen
2018-05-22 14:37:27 +0200
c54f7aca24
Merge branch 'experimentation' into 'master'
Christoffer Müller Madsen
2018-05-22 12:36:37 +0000
c31bc39780
More server
Alexander Munch-Hansen
2018-05-22 00:26:32 +0200
6133cb439f
Merge remote-tracking branch 'origin/experimentation' into experimentation
Alexander Munch-Hansen
2018-05-20 20:15:57 +0200
5acd79b6da
Slight modification to move calculation
Alexander Munch-Hansen
2018-05-20 19:43:28 +0200
b11e783b30
add 0-ply-tests
2018-05-20 18:50:28 +0200
f834b10e02
remove unnecessary print
Christoffer Müller Madsen
2018-05-20 16:52:05 +0200
72f01a2a2d
remove dependency on yaml
Christoffer Müller Madsen
2018-05-20 16:03:58 +0200
d14e6c5994
Everything might work, except for quad, that might be bugged.
Alexander Munch-Hansen
2018-05-20 00:38:13 +0200
a266293ecd
Stuff is happening, moving is better!
Alexander Munch-Hansen
2018-05-19 22:01:55 +0200
e9a46c79df
server and stuff
Alexander Munch-Hansen
2018-05-19 14:12:13 +0200
816cdfae00
fix and clean
Alexander Munch-Hansen
2018-05-18 14:55:10 +0200
ff9664eb38
Merge branch 'eager_eval' into 'master'
Christoffer Müller Madsen
2018-05-18 12:06:12 +0000
3e379b40c4
Accidentally added a '5' in the middle of a variable.
Alexander Munch-Hansen
2018-05-16 00:20:54 +0200
90fad334b9
More optimizations.
Alexander Munch-Hansen
2018-05-15 23:37:35 +0200
a77c13a0a4
1-ply runs even faster.
Alexander Munch-Hansen
2018-05-15 19:29:27 +0200
260c32d909
oiuhhiu
Alexander Munch-Hansen
2018-05-15 18:16:44 +0200
00974b0f11
Added '--play' flag, so you can now play against the ai.
Alexander Munch-Hansen
2018-05-14 13:07:48 +0200
2c02689577
Merge remote-tracking branch 'origin/eager_eval' into eager_eval
Alexander Munch-Hansen
2018-05-13 23:55:02 +0200
926a331df0
Some flags from main.py is gone, rolls now allow a face_value of 0 yet again and it is possible to play against the ai. There is no flag for this yet, so this has to be added.
Alexander Munch-Hansen
2018-05-13 23:54:13 +0200
d932663519
add explanation of ply speedup
Christoffer Müller Madsen
2018-05-13 22:26:24 +0200
2312c9cb2a
Merge branch 'eager_eval' of gitfub.space:Pownie/backgammon into eager_eval
Christoffer Müller Madsen
2018-05-12 15:19:12 +0200
9f1bd56c0a
fix bear_off bug; addtional tests and additional fixes
Christoffer Müller Madsen
2018-05-12 15:18:52 +0200
ba4ef86bb5
Board rep can now be inferred from file after being given once. We can also evaluate multiple times by using the flag "--repeat-eval". The flag defaults to 1, if not provided.
Alexander Munch-Hansen
2018-05-12 12:14:47 +0200
c3f5e909d6
flip is back
Christoffer Müller Madsen
2018-05-11 21:47:48 +0200
1aa9cf705f
quack without leaks
Christoffer Müller Madsen
2018-05-11 21:24:10 +0200
383dd7aa4b
code works again; quack gave ~3 times improvement for calc_moves
Christoffer Müller Madsen
2018-05-11 20:13:43 +0200
93188fe06b
more quack for board
Christoffer Müller Madsen
2018-05-11 20:07:27 +0200
ffbc98e1a2
quack kind of works
Christoffer Müller Madsen
2018-05-11 19:00:39 +0200
03e61a59cf
quack
Christoffer Müller Madsen
2018-05-11 17:29:22 +0200
93224864a4
More comments, backprop have been somewhat tested in the eager_main.py and normal_main.py.
Alexander Munch-Hansen
2018-05-11 13:35:01 +0200
504308a9af
Yet another input argument, "--ply", 0 for no look-ahead, 1 for a single look-ahead.
Alexander Munch-Hansen
2018-05-10 23:22:41 +0200
3b57c10b5a
Saves calling tf.reduce_mean on all values once.
Alexander Munch-Hansen
2018-05-10 22:57:27 +0200
4fa10861bb
update TF dependency to 1.8.0
Christoffer Müller Madsen
2018-05-10 19:27:51 +0200
6131d5b5f4
Added comments for Christoffer!
Alexander Munch-Hansen
2018-05-10 19:25:28 +0200
1aedc23de1
1-ply now works again.
Alexander Munch-Hansen
2018-05-10 19:13:18 +0200
2d84cd5a0b
1-ply now works again.
Alexander Munch-Hansen
2018-05-10 19:06:53 +0200
396d5b036d
All values for boards and all rolls can now be calculated
Alexander Munch-Hansen
2018-05-10 18:41:21 +0200
4efb229d34
Added a lot of comments
Alexander Munch-Hansen
2018-05-10 15:28:33 +0200
f2a67ca92e
All board reps should now work as input.
Alexander Munch-Hansen
2018-05-10 10:49:25 +0200
9cfdd7e2b2
Added a verbosity flag, --verbose, which allows for printing of variables and such.
Alexander Munch-Hansen
2018-05-10 10:39:22 +0200
6429e0732c
We should now be able to both train and eval as per usual. I've added a file "global_step", which works as the new global_step counter, so we can use it for exp_decay.
Alexander Munch-Hansen
2018-05-09 23:15:35 +0200
cb7e7b519c
Getting closer to functionality. We're capable of evaluating moves and a rework of global_step has begun, such that we now use episode_count as a way of calculating exp_decay, which have been implemented as a function.
Alexander Munch-Hansen
2018-05-09 22:22:12 +0200
9a2d87516e
Ongoing rewrite of network to use an eager model. We're now capable of evaluating a list of states with network.py. We can also save and restore models.
Alexander Munch-Hansen
2018-05-09 00:33:05 +0200
7b308be4e2
Different implementations of different speed
Alexander Munch-Hansen
2018-05-07 22:24:47 +0200
ac6660e05b
Added board-rep as cli argument, to state which input-board-rep to use. Also fixed weird nesting of difference_in_values.
Alexander Munch-Hansen
2018-05-06 20:52:35 +0200
1f8485f54e
No longer use n_ply, shit's too slow man. Added extra logging, now logs the average difference in values between trainings. Also fixed bug with the length of quack-norm. Also added cli argument; use-baseline, if set, the baseline-model will be used.
Alexander Munch-Hansen
2018-05-06 20:41:07 +0200
1db469709a
make_move now calls n_ply to search deeper and potentially give better moves. It's hella fucking slow.
Alexander Munch-Hansen
2018-05-02 01:06:23 +0200
695a3d43db
Fixed n_ply and actually added a comma in main.py. *clap Christoffer*
Alexander Munch-Hansen
2018-05-01 20:39:29 +0200
c530aa688d
flipidip
Christoffer Müller Madsen
2018-05-01 13:48:42 +0200
3f6849048e
added network_test and some comments
Alexander Munch-Hansen
2018-04-29 12:14:14 +0200
afa6504b05
ply again again
Christoffer Müller Madsen
2018-04-26 16:49:49 +0200
9428a00c11
add "--force-creation" flag to force model creation
Christoffer Müller Madsen
2018-04-26 11:43:19 +0200
48a5f6cbb6
Moved "do_ply" out of "calculate_2_ply", in an effort to be able to eventually do further plies, however some rewriting of the current "do_ply" will be needed, as described in a comment.
Pownie
2018-04-26 09:42:03 +0200
8899c5c2d9
Fixed potential bug in regards to scores in 2-ply calculation.
Pownie
2018-04-25 00:51:04 +0200
ea3f05846d
Merge branch 'master' of https://gitfub.space/Pownie/backgammon
Pownie
2018-04-24 22:31:18 +0200
0509a51fd3
Added baseline model for testing
Pownie
2018-04-24 22:30:58 +0200
33a4b0db3c
disallow using model "baseline"
Christoffer Müller Madsen
2018-04-24 21:16:54 +0200
349ad718f1
Moved gen_21_rolls into the 2-ply method, so it can be correctly used like the good helper method that it is
Pownie
2018-04-23 00:45:31 +0200
e5cc54d3e0
Added a normalised version of quack
Pownie
2018-04-23 00:35:25 +0200
160f5bd737
added some comments and removed some old code
Pownie
2018-04-22 19:13:46 +0200
77d82f6883
Added code for 2-ply look-ahead
Pownie
2018-04-22 15:07:19 +0200
1062b72bda
fix typo
Christoffer Müller Madsen
2018-04-19 16:04:49 +0200
66589dfde3
fixed global step, now using exp decay
Alexander Munch-Hansen
2018-04-19 16:01:19 +0200
cba0f67ae2
fixed *the* bug
Alexander Munch-Hansen
2018-04-19 15:22:00 +0200
b6c52ba476
fix type error
Christoffer Müller Madsen
2018-04-16 00:24:24 +0200
8998dca1f2
remove @Pownie 's debug print
Christoffer Müller Madsen
2018-04-16 00:03:02 +0200
611f6cdba0
Changed alpha to learning_rate
Pownie
2018-04-15 23:53:35 +0200
57fb1cb141
Merge branch 'master' of https://gitfub.space/Pownie/backgammon
Pownie
2018-04-15 23:52:00 +0200
cc1e010840
Uses proper board instead of Alex' drunken mistakes
Pownie
2018-04-15 23:51:28 +0200
f68d7a9ded
add pygame to requirements.txt
Christoffer Müller Madsen
2018-04-15 22:45:37 +0200
f59fe27e5f
You can now move off bar
Pownie
2018-04-14 23:31:33 +0200
7d29fc02f2
Added global step + exponential decay
Pownie
2018-04-14 23:11:20 +0200
1d9c94896d
Red can go on bar as well now
Pownie
2018-04-14 22:53:49 +0200
716413e2b6
bar works somewhat if black goes on there. Still can't get off it
Pownie
2018-04-14 22:51:41 +0200
7993da0db7
Turns are now functioning
Pownie
2018-04-14 18:47:38 +0200
7764a70799
Changed calculate_legal_states to allow for possible face_value of 0
Alexander Munch-Hansen
2018-04-14 14:51:50 +0200
Christoffer Müller Madsen
Alexander Munch-Hansen
=