-
Notifications
You must be signed in to change notification settings - Fork 0
/
qlearn.py
140 lines (115 loc) · 4.08 KB
/
qlearn.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
#!/usr/bin/env python
# -*- coding: utf-8 -*-
import random
import gym
import universe
import numpy as np
import math
from replaymem import ReplayMemory, Transition
from keras.optimizers import RMSprop
from keras.models import Sequential
from keras.layers import Conv3D
from keras.layers import Dense, Dropout, Flatten
GAME = 'flashgames.NeonRace2-v0'
LEFT = [('KeyEvent', 'ArrowLeft', True),('KeyEvent', 'ArrowRight', False),('KeyEvent', 'ArrowUp', True),('KeyEvent', 'ArrowDown', False)]
RIGHT = [('KeyEvent', 'ArrowLeft', False),('KeyEvent', 'ArrowRight', True),('KeyEvent', 'ArrowUp', True),('KeyEvent', 'ArrowDown', False)]
FORWARD = [('KeyEvent', 'ArrowLeft', False),('KeyEvent', 'ArrowRight', False),('KeyEvent', 'ArrowUp', True),('KeyEvent', 'ArrowDown', False)]
ACTIONS = [LEFT, RIGHT, FORWARD]
BATCH_SIZE = 64
GAMMA = 0.999
EPS_START = 0.9
EPS_END = 0.05
EPS_DECAY = 200
memory = ReplayMemory(10000)
done_n = [False]
steps_done = 0
Q = Sequential()
Q.add(Conv3D(32, kernel_size=(5,5,3), strides=5,
activation='relu',
input_shape=(768, 1024, 3, 1)))
Q.add(Conv3D(64, kernel_size=(5,5,1), strides=3,
activation='relu',
input_shape=(255, 340, 1, 32)))
Q.add(Conv3D(64, kernel_size=(4,4,1), strides=2,
activation='relu',
input_shape=(50, 67, 1, 64)))
Q.add(Flatten())
Q.add(Dense(1000, activation='relu', input_shape=(49152,)))
Q.add(Dense(512, activation='relu', input_shape=(1000,)))
Q.add(Dense(128, activation='relu', input_shape=(512,)))
Q.add(Dense(3, activation='linear', input_shape=(128,)))
Q.compile(loss='categorical_crossentropy',
optimizer=RMSprop(),
metrics=['accuracy'])
def select_action(state):
global steps_done
global Q
sample = random.random()
eps_threshold = EPS_END + (EPS_START - EPS_END) * \
math.exp(-1. * steps_done / EPS_DECAY)
steps_done += 1
if sample > eps_threshold:
return np.argmax(Q.predict(state))
else:
return random.randint(0,2)
def optimize_model():
print('Starting optimization')
if len(memory) < BATCH_SIZE:
return
transitions = memory.sample(BATCH_SIZE)
batch = Transition(*zip(*transitions))
x_train = []
y_train = []
for i in range(len(batch[0])):
ss = batch[0][i]
aa = batch[1][i]
if(batch[2][i] != None):
ss_p = np.expand_dims(np.expand_dims(batch[2][i][0]['vision'], axis=3), axis=0)
rr = batch[3][i]
tt = rr
if(batch[2][i] != None):
tt = rr + GAMMA*np.amax(Q.predict(ss_p))
x_train.append(ss)
y_data = np.zeros(3)
y_data[aa] = tt
y_train.append(y_data)
x_train = np.array(x_train)
y_train = np.array(y_train)
print(x_train)
print(y_train)
Q.fit(x=x_train, y=y_train, batch_size=2, epochs=2, verbose=1)
episodes = 5
def main():
env = gym.make(GAME)
env.configure(remotes='vnc://localhost:5900+15900')
# env.configure(remotes=1)
observation_n = env.reset()
e = 0
while(e <= episodes):
if(observation_n[0] != None):
e+=1
print('Entered')
for time_t in range(500):
observation = np.expand_dims(observation_n[0]['vision'], axis=3)
action = select_action(np.expand_dims(observation, axis=0))
next_state, reward_n, done_n, info = env.step([ACTIONS[action]])
env.render()
if(done_n[0]):
memory.ins(observation, action, [None], reward_n)
print("episode: {}/{}, score: {}"
.format(e, episodes, time_t))
observation_n = env.reset()
break
else:
memory.ins(observation, action, next_state, reward_n)
else:
print("Game Restart")
observation_n = env.reset()
optimize_model()
else:
action_n = [random.choice(ACTIONS)]
next_state, reward_n, done_n, info = env.step(action_n)
observation_n = next_state
env.render()
if __name__ == '__main__':
main()