Map=\
['#o#*#o#*#',
'o1o *o^o',
'# # # #o#',
'o #####^o',
'#*#####o#',
'oo^oo^ o',
'#oo^#oo0#'
]
human_play=False
have_render=True
env = Maze(Map)
if(have_render):
Image = env.render()
act=[0,0]
score=[0,0]
def ch(x):
if(x=='l'):return 0
if(x=='r'):return 1
if(x=='u'):return 2
if(x=='d'):return 3
if(x=='b'):return 4
if(x=='s'):return 5
def read_action(event):
if event.keysym == 'Left':
act[0]='l'
if event.keysym == 'Right':
GLOBAL_UPDATE_COUNTER, GLOBAL_EP = 0, 0
GLOBAL_RUNNING_R = []
COORD = tf.train.Coordinator()
QUEUE = queue.Queue() # workers putting data in this queue
threads = []
for worker in workers: # worker threads
t = threading.Thread(target=worker.work, args=())
t.start() # training
threads.append(t)
# add a PPO updating thread
threads.append(threading.Thread(target=GLOBAL_PPO.update, ))
threads[-1].start()
COORD.join(threads)
print('aasdas')
env = Maze(Map)
tf.reset_default_graph()
load_PPO = PPO(Load=True)
while True:
s = env.reset()
for t in range(100):
env.render()
a = load_PPO.choose_action(s)
baseline_a = base.choose_action(env, 1)
s, r, done = env.step({(0, a), (1, U.ch(baseline_a))})
if (r[0] != -1):
print(r)
if (done):
break