Max recursion depth exceeded

[brianprichardson]

Othello TF per below on iteration 1, epoch 10.
BTW: With TF less verbosity would be helpful (400 lines per epoch)

Training Net |############################### | (400/403) Data: 0.000s | Batch: 0.031s | Total: 0:00:12 | ETA: 0:00:01 | Loss_pi: 3.4790 |Training Net |############################### | (401/403) Data: 0.000s | Batch: 0.031s | Total: 0:00:12 | ETA: 0:00:01 | Loss_pi: 3.4789 |Training Net |############################### | (402/403) Data: 0.000s | Batch: 0.031s | Total: 0:00:12 | ETA: 0:00:01 | Loss_pi: 3.4789 |Training Net |################################| (403/403) Data: 0.000s | Batch: 0.031s | Total: 0:00:12 | ETA: 0:00:01 | Loss_pi: 3.4789 | Loss_v: 0.382
PITTING AGAINST PREVIOUS VERSION
/home/brian/hitme/bin/alpha-zero-general/MCTS.py:80: RuntimeWarning: invalid value encountered in true_divide
self.Ps[s] /= np.sum(self.Ps[s]) # renormalize
Traceback (most recent call last):
File "main.py", line 29, in
c.learn()
File "/home/brian/hitme/bin/alpha-zero-general/Coach.py", line 99, in learn
pwins, nwins, draws = arena.playGames(self.args.arenaCompare)
File "/home/brian/hitme/bin/alpha-zero-general/Arena.py", line 81, in playGames
gameResult = self.playGame(verbose=verbose)
File "/home/brian/hitme/bin/alpha-zero-general/Arena.py", line 46, in playGame
action = players[curPlayer+1](self.game.getCanonicalForm(board, curPlayer))
File "/home/brian/hitme/bin/alpha-zero-general/Coach.py", line 98, in
lambda x: np.argmax(nmcts.getActionProb(x, temp=0)), self.game)
File "/home/brian/hitme/bin/alpha-zero-general/MCTS.py", line 31, in getActionProb
self.search(canonicalBoard)
File "/home/brian/hitme/bin/alpha-zero-general/MCTS.py", line 106, in search
v = self.search(next_s)
File "/home/brian/hitme/bin/alpha-zero-general/MCTS.py", line 106, in search
v = self.search(next_s)
File "/home/brian/hitme/bin/alpha-zero-general/MCTS.py", line 106, in search
v = self.search(next_s)
[Previous line repeated 983 more times]
File "/home/brian/hitme/bin/alpha-zero-general/MCTS.py", line 103, in search
next_s, next_player = self.game.getNextState(canonicalBoard, 1, a)
File "/home/brian/hitme/bin/alpha-zero-general/othello/OthelloGame.py", line 31, in getNextState
b = Board(self.n)
File "/home/brian/hitme/bin/alpha-zero-general/othello/OthelloLogic.py", line 24, in init
for i in range(self.n):
RecursionError: maximum recursion depth exceeded in comparison

[evg-tyurin]

Most likely game logic error leading to infinite loop in traversing position tree. Game over condition can't be detected.

[suragnair]

Would it be possible to have it print the state that causes this?

[brianprichardson]

Any suggestions as to how to do so?

[evg-tyurin]

You can catch the error by surrounding
next_s, next_player = self.game.getNextState(canonicalBoard, 1, a)
with try / except clause and in except part print the state or anything you want.

[emc2314]

I've encountered the same problem. So I display the board using try / except. Here are some examples:

0 |1 |2 |3 |4 |5 |
 -----------------------
0 |b b b b b b |
1 |b b b b b b |
2 |W W b b b b |
3 |b b b b b b |
4 |b b W b - b |
5 |b b b W - - |
 -----------------------
0 |b b b b b b |
1 |b b b W b b |
2 |b b W b W b |
3 |b b b b b b |
4 |b b b b b b |
5 |b b b b - b |
 ------------------------
0 |b - b W W W |
1 |b b b b b W |
2 |b b b b b b |
3 |b b b b b b |
4 |b b b - b b |
5 |b b b b b b |
   -----------------------

[emc2314]

And I decided to print states after the exception. Oops!

   -----------------------
0 |1 |2 |3 |4 |5 |
 -----------------------
0 |b b W W W W |
1 |b b b b b W |
2 |b b W b b W |
3 |b b b b - b |
4 |b b W b b b |
5 |b b b b b b |
   -----------------------
0 |1 |2 |3 |4 |5 |
 -----------------------
0 |W W b b b b |
1 |W W W W W b |
2 |W W b W W b |
3 |W W W W - W |
4 |W W b W W W |
5 |W W W W W W |
   -----------------------
0 |1 |2 |3 |4 |5 |
 -----------------------
0 |b b W W W W |
1 |b b b b b W |
2 |b b W b b W |
3 |b b b b - b |
4 |b b W b b b |
5 |b b b b b b |
   -----------------------
0 |1 |2 |3 |4 |5 |
 -----------------------
0 |W W b b b b |
1 |W W W W W b |
2 |W W b W W b |
3 |W W W W - W |
4 |W W b W W W |
5 |W W W W W W |
   -----------------------
...

[suragnair]

@emc2314 Are the above snapshots from the same game? Could you share your exact code with the try except?

[emc2314]

Problem confirmed:
In Line 96 of MCTS.py, self.Ps[s][a] is nan, resulting a = -1, so the game cannot end.

My code is really kind of rubbish... I feel ashamed to put it here.
And the above snapshots are all from one checkpoint, but the problem dose not always come out so I have to try many times to debug. Sorry for the delay when reply.

[suragnair]

This should be linked to #23.

[evg-tyurin]

Related to the game logic, I'd propose to add the move validation into board.execute_move(). If the move supplied by an invoker is invalid for the current board position an error should be raised.

I mean the rule that the player must not pass the move (a=-1) if he has one valid.

[emc2314]

Actually there is an assertion.

alpha-zero-general/othello/OthelloLogic.py

Line 111 in 2a37807

assert len(list(flips))>0

Sometimes it works, but sometimes it doesn't..

[suragnair]

Addressed in f67e87e .