add multi-thread for end-to-end training

This commit is contained in:
rtz19970824 2018-01-13 15:57:41 +08:00
parent fcaa571b42
commit 2e8662889f
2 changed files with 81 additions and 73 deletions

View File

@ -119,13 +119,12 @@ class ResNet(object):
zip(self.black_var_list, self.white_var_list)]
# training hyper-parameters:
self.window_length = 900
self.window_length = 500
self.save_freq = 5000
self.training_data = {'states': deque(maxlen=self.window_length), 'probs': deque(maxlen=self.window_length),
'winner': deque(maxlen=self.window_length), 'length': deque(maxlen=self.window_length)}
# training or not
self.training = False
self.use_latest = False
def _build_network(self, scope, residual_block_num):
"""
@ -188,16 +187,16 @@ class ResNet(object):
feed_dict={self.x: eval_state, self.is_training: False})
def check_latest_model(self):
if self.training:
if self.use_latest:
black_ckpt_file = tf.train.latest_checkpoint(self.save_path + "black/")
if self.black_ckpt_file != black_ckpt_file:
if self.black_ckpt_file != black_ckpt_file and black_ckpt_file is not None:
self.black_ckpt_file = black_ckpt_file
print('Loading model from {}...'.format(self.black_ckpt_file))
self.black_saver.restore(self.sess, self.black_ckpt_file)
print('Black Model Updated!')
white_ckpt_file = tf.train.latest_checkpoint(self.save_path + "white/")
if self.white_ckpt_file != white_ckpt_file:
if self.white_ckpt_file != white_ckpt_file and white_ckpt_file is not None:
self.white_ckpt_file = white_ckpt_file
print('Loading model from {}...'.format(self.white_ckpt_file))
self.white_saver.restore(self.sess, self.white_ckpt_file)
@ -234,7 +233,7 @@ class ResNet(object):
:param target: a string, which to optimize, can only be "both", "black" and "white"
:param mode: a string, how to optimize, can only be "memory" and "file"
"""
self.training = True
self.use_latest = True
if mode == 'memory':
pass
if mode == 'file':
@ -401,5 +400,5 @@ class ResNet(object):
if __name__ == "__main__":
model = ResNet(board_size=8, action_num=65, history_length=1, black_checkpoint_path="./checkpoint/black", white_checkpoint_path="./checkpoint/white")
model.train(mode="file", data_path="./data/", batch_size=128, save_path="./checkpoint/")
model = ResNet(board_size=9, action_num=82, history_length=8, black_checkpoint_path="./checkpoint/black", white_checkpoint_path="./checkpoint/white")
model.train(mode="file", data_path="./data/", batch_size=128, save_path="./go-v2/")

View File

@ -3,6 +3,7 @@ import sys
import re
import time
import os
import threading
from game import Game
from engine import GTPEngine
from utils import Data
@ -17,6 +18,67 @@ else:
import _pickle as cPickle
def play(engine, data_path):
data = Data()
role = ["BLACK", "WHITE"]
color = ['b', 'w']
pattern = "[A-Z]{1}[0-9]{1}"
space = re.compile("\s+")
size = {"go": 9, "reversi": 8}
show = ['.', 'X', 'O']
# evaluate_rounds = 100
game_num = 0
while True:
# while game_num < evaluate_rounds:
engine._game.model.check_latest_model()
num = 0
pass_flag = [False, False]
print("Start game {}".format(game_num))
# end the game if both palyer chose to pass, or play too much turns
while not (pass_flag[0] and pass_flag[1]) and num < size[engine._game.name] ** 2 * 2:
turn = num % 2
board = engine.run_cmd(str(num) + ' show_board')
board = eval(board[board.index('['):board.index(']') + 1])
for i in range(size[engine._game.name]):
for j in range(size[engine._game.name]):
print show[board[i * size[engine._game.name] + j]] + " ",
print "\n",
data.boards.append(board)
move = engine.run_cmd(str(num) + ' genmove ' + color[turn])[:-1]
print("\n" + role[turn] + " : " + str(move)),
num += 1
match = re.search(pattern, move)
if match is not None:
# print "match : " + str(match.group())
pass_flag[turn] = False
else:
# print "no match"
pass_flag[turn] = True
prob = engine.run_cmd(str(num) + ' get_prob')
prob = space.sub(',', prob[prob.index('['):prob.index(']') + 1])
prob = prob.replace('[,', '[')
prob = prob.replace('],', ']')
prob = eval(prob)
data.probs.append(prob)
score = engine.run_cmd(str(num) + ' get_score')
print("Finished : {}".format(score.split(" ")[1]))
if eval(score.split(" ")[1]) > 0:
data.winner = utils.BLACK
if eval(score.split(" ")[1]) < 0:
data.winner = utils.WHITE
engine.run_cmd(str(num) + ' clear_board')
current_time = strftime("%Y%m%d_%H%M%S", gmtime())
if os.path.exists(data_path + current_time + ".pkl"):
time.sleep(1)
current_time = strftime("%Y%m%d_%H%M%S", gmtime())
with open(data_path + current_time + ".pkl", "wb") as file:
cPickle.dump(data, file)
data.reset()
game_num += 1
if __name__ == '__main__':
"""
Starting two different players which load network weights to evaluate the winning ratio.
@ -27,6 +89,7 @@ if __name__ == '__main__':
parser.add_argument("--data_path", type=str, default="./data/")
parser.add_argument("--black_weight_path", type=str, default=None)
parser.add_argument("--white_weight_path", type=str, default=None)
parser.add_argument("--save_path", type=str, default="./go/")
parser.add_argument("--debug", type=bool, default=False)
parser.add_argument("--game", type=str, default="go")
args = parser.parse_args()
@ -46,69 +109,15 @@ if __name__ == '__main__':
debug=args.debug)
engine = GTPEngine(game_obj=game, name='tianshou', version=0)
data = Data()
role = ["BLACK", "WHITE"]
color = ['b', 'w']
thread_list = []
thread_train = threading.Thread(target=game.model.train, args=("file",),
kwargs={'data_path':args.data_path, 'batch_size':128, 'save_path':args.save_path})
thread_play = threading.Thread(target=play, args=(engine, args.data_path))
thread_list.append(thread_train)
thread_list.append(thread_play)
pattern = "[A-Z]{1}[0-9]{1}"
space = re.compile("\s+")
size = {"go":9, "reversi":8}
show = ['.', 'X', 'O']
for t in thread_list:
t.start()
evaluate_rounds = 100
game_num = 0
try:
while True:
#while game_num < evaluate_rounds:
start_time = time.time()
game.model.check_latest_model()
num = 0
pass_flag = [False, False]
print("Start game {}".format(game_num))
# end the game if both palyer chose to pass, or play too much turns
while not (pass_flag[0] and pass_flag[1]) and num < size[args.game] ** 2 * 2:
turn = num % 2
board = engine.run_cmd(str(num) + ' show_board')
board = eval(board[board.index('['):board.index(']') + 1])
for i in range(size[args.game]):
for j in range(size[args.game]):
print show[board[i * size[args.game] + j]] + " ",
print "\n",
data.boards.append(board)
start_time = time.time()
move = engine.run_cmd(str(num) + ' genmove ' + color[turn])[:-1]
print("\n" + role[turn] + " : " + str(move)),
num += 1
match = re.search(pattern, move)
if match is not None:
# print "match : " + str(match.group())
play_or_pass = match.group()
pass_flag[turn] = False
else:
# print "no match"
play_or_pass = ' PASS'
pass_flag[turn] = True
prob = engine.run_cmd(str(num) + ' get_prob')
prob = space.sub(',', prob[prob.index('['):prob.index(']') + 1])
prob = prob.replace('[,', '[')
prob = prob.replace('],', ']')
prob = eval(prob)
data.probs.append(prob)
score = engine.run_cmd(str(num) + ' get_score')
print("Finished : {}".format(score.split(" ")[1]))
if eval(score.split(" ")[1]) > 0:
data.winner = utils.BLACK
if eval(score.split(" ")[1]) < 0:
data.winner = utils.WHITE
engine.run_cmd(str(num) + ' clear_board')
file_list = os.listdir(args.data_path)
current_time = strftime("%Y%m%d_%H%M%S", gmtime())
if os.path.exists(args.data_path + current_time + ".pkl"):
time.sleep(1)
current_time = strftime("%Y%m%d_%H%M%S", gmtime())
with open(args.data_path + current_time + ".pkl", "wb") as file:
picklestring = cPickle.dump(data, file)
data.reset()
game_num += 1
except KeyboardInterrupt:
pass
for t in thread_list:
t.join()