def play(network):
"""Plays out a self-play match, returning a MCTSPlayer object containing:
- the final position
- the n x 362 tensor of floats representing the mcts search probabilities
- the n-ary tensor of floats representing the original value-net estimate
where n is the number of moves in the game
"""
readouts = FLAGS.num_readouts # defined in strategies.py
# Disable resign in 5% of games
if random.random() < FLAGS.resign_disable_pct:
resign_threshold = -1.0
else:
resign_threshold = None
player = MCTSPlayer(network, resign_threshold=resign_threshold)
player.initialize_game()
# Must run this once at the start to expand the root node.
first_node = player.root.select_leaf()
prob, val = network.run(first_node.position)
first_node.incorporate_results(prob, val, first_node)
while True:
start = time.time()
player.root.inject_noise()
current_readouts = player.root.N
# we want to do "X additional readouts", rather than "up to X readouts".
while player.root.N < current_readouts + readouts:
player.tree_search()
if FLAGS.verbose >= 3:
print(player.root.position)
print(player.root.describe())
if player.should_resign():
player.set_result(-1 * player.root.position.to_play,
was_resign=True)
break
move = player.pick_move()
player.play_move(move)
if player.root.is_done():
player.set_result(player.root.position.result(), was_resign=False)
break
if (FLAGS.verbose >= 2) or (FLAGS.verbose >= 1 and player.root.position.n % 10 == 9):
print("Q: {:.5f}".format(player.root.Q))
dur = time.time() - start
print("%d: %d readouts, %.3f s/100. (%.2f sec)" % (
player.root.position.n, readouts, dur / readouts * 100.0, dur), flush=True)
if FLAGS.verbose >= 3:
print("Played >>",
coords.to_gtp(coords.from_flat(player.root.fmove)))
if FLAGS.verbose >= 2:
utils.dbg("%s: %.3f" % (player.result_string, player.root.Q))
utils.dbg(player.root.position, player.root.position.score())
return player
def play(network):
search_n = 100
player = MCTSPlayer(network=network,seconds_per_move=seconds_per_move,timed_match=timed_match,search_n=search_n, player_mode=0)
player.initialize_game()
while True:
start = time.time()
current_n = player.root.N
while player.root.N < current_n + search_n:
player.tree_search()
move = player.pick_move()
#print(move, player.root.status.to_play)
player.play_move(move)
if player.root.is_done():
#print('[!] finish')
break
#X, p, v = player.generate_data()
return player
def play(network):
readouts = FLAGS.num_readouts
player = MCTSPlayer(network)
player.initialize_game()
first_node = player.root.select_leaf()
prob, val = network.predict(first_node.position.state)
first_node.incorporate_results(prob, val, first_node)
while True:
# player.root.inject_noise()
current_readouts = player.root.N
while player.root.N < current_readouts + readouts:
player.tree_search()
move = player.pick_move()
player.play_move(move)
tf.logging.info('playing move: %d hamming distance: %d' % (move, state_diff(player.root.position.state)))
if player.root.is_done():
tf.logging.info('done')
break
def play(network):
readouts = FLAGS.num_readouts
player = MCTSPlayer(network)
player.initialize_game()
first_node = player.root.select_leaf()
prob, val = network.predict(first_node.position.state)
first_node.incorporate_results(prob, val, first_node)
lastmove = -1
hamm_dist = state_diff(player.root.position.state)
for lo in range(0, hamm_dist):
# player.root.inject_noise()
current_readouts = player.root.N
start = time.time()
while player.root.N < current_readouts + readouts and time.time(
) - start < FLAGS.time_per_move:
player.tree_search()
move = player.pick_move()
if move == lastmove:
tf.logging.info('lastmove == move')
return state_diff(player.root.position.state)
before = state_diff(player.root.position.state)
player.play_move(move)
after = state_diff(player.root.position.state)
if after > before:
tf.logging.info('move increasing distance')
return after
tf.logging.info('playing move: %d hamming distance: %d' %
(move, state_diff(player.root.position.state)))
if player.root.is_done():
tf.logging.info('done')
return 0
lastmove = move
return state_diff(player.root.position.state)
