-
Notifications
You must be signed in to change notification settings - Fork 0
/
a2cAgent.py
244 lines (210 loc) · 11 KB
/
a2cAgent.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
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
import torch
import numpy as np
import torch.nn as nn
import cv2
import random
import os
import datetime
import argparse
from torch.nn import functional as F
from torch.utils.tensorboard import SummaryWriter
from game.doodlejump import DoodleJump
from model.networks import ActorCritic
from model.a2cTrainer import A2CLearner
def mish(input):
return input * torch.tanh(F.softplus(input))
class Mish(nn.Module):
def __init__(self): super().__init__()
def forward(self, input): return mish(input)
def t(x):
x = np.array(x) if not isinstance(x, np.ndarray) else x
return torch.from_numpy(x).float() #.reshape(6400, 1)
class Runner():
def __init__(self, game):
self.game = game
self.state = None
self.done = True
self.steps = 0
self.episode_reward = 0
self.mean_reward = 0
self.mean_score = 0
self.episode_rewards = []
self.total_score = 0
''''''
self.n_games = 0
self.epsilon = 0
self.ctr = 1
seed = args.seed
os.environ['PYTHONHASHSEED'] = str(seed)
# Torch RNG
torch.manual_seed(seed)
torch.cuda.manual_seed(seed)
torch.cuda.manual_seed_all(seed)
# Python RNG
np.random.seed(seed)
random.seed(seed)
self.store_frames = args.store_frames
self.image_h = args.height
self.image_w = args.width
self.image_c = args.channels
# self.memory = deque(maxlen=args.max_memory)
self.device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
self.gamma = args.gamma
self.batch_size = args.batch_size
self.record = 0
def reset(self):
self.episode_reward = 0
self.done = False
self.state = None
self.game.gameReboot()
def preprocess(self, state):
# resize the image and then rotate
img = cv2.resize(state, (self.image_w, self.image_h))
M = cv2.getRotationMatrix2D((self.image_w / 2, self.image_h / 2), 270, 1.0)
img = cv2.warpAffine(img, M, (self.image_h, self.image_w))
if self.store_frames:
os.makedirs("./image_dump", exist_ok=True)
cv2.imwrite("./image_dump/"+str(self.ctr)+".jpg", img)
self.ctr+=1
imagenet_mean = [0.485, 0.456, 0.406]
imagenet_std = [0.229, 0.224, 0.225]
if self.image_c == 1:
# convert the image to grayscale
img = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)
# normalize the image with imagenet mean and std values
img = ((img/255.0) - np.mean(imagenet_mean))/np.mean(imagenet_std)
else:
# normalize the image with imagenet mean and std values
img = ((img/255.0) - imagenet_mean)/imagenet_std
# change the shape from WxHxC to CxHxW for pytorch tensor
img = img.transpose((2, 0, 1))
# Add a axis for converting image to shape: 1xCxHxW
img = np.expand_dims(img, axis=0)
return img
def get_state(self):
state = self.game.getCurrentFrame()
return self.preprocess(state)
def run(self, max_steps, memory=None):
loop_ctr = 0
if not memory or len(memory) > args.max_memory:
memory = []
for _ in range(max_steps):
if self.done: self.reset()
state_old = self.get_state()
dists = actorcritic(t(state_old).to(self.device))[0]
actions = dists.sample().detach().cpu().data.numpy()
actions_clipped = np.clip(actions, -1, 1) #self.env.action_space.low.min(), env.action_space.high.max())
final_move = [0,0,0]
final_move[np.argmax(actions_clipped)] = 1
reward, self.done, score = self.game.playStep(final_move)
next_state = self.get_state()
memory.append((actions, reward, self.state, next_state, self.done))
self.state = next_state
self.steps += 1
self.episode_reward += reward
self.mean_reward = self.episode_reward/self.steps
writer.add_scalar("Reward/mean_reward", self.mean_reward, global_step=self.steps)
if self.done:
self.n_games += 1
learner.lastlaegam = 0
self.episode_rewards.append(self.episode_reward)
if len(self.episode_rewards) % 10 == 0:
print("episode:", len(self.episode_rewards), ", episode reward:", self.episode_reward)
writer.add_scalar("Reward/episode_reward", self.episode_reward, global_step=self.steps)
if score > self.record:
self.record = score
# save the best model yet
actorcritic.save(file_name="a2c_model_best.pth", model_folder_path="./model"+hyper_params+dstr)
# actorcritic.save(file_name="critic_model_best.pth", model_folder_path="./model"+hyper_params+dstr)
if self.n_games%100 == 0:
# save model per 100 games
actorcritic.save(file_name="a2c_model_"+str(self.n_games)+".pth", model_folder_path="./model"+hyper_params+dstr)
# actorcritic.save(file_name="critic_model_"+str(self.n_games)+".pth", model_folder_path="./model"+hyper_params+dstr)
print('Game', self.n_games, 'Score', score, 'Record:', self.record)
writer.add_scalar('Score/High_Score', self.record, self.n_games)
self.total_score += score
self.mean_score = self.total_score / agent.n_games
writer.add_scalar('Score/Mean_Score', self.mean_score, self.n_games)
return memory
def test(game, args):
if args.macos:
os.environ['KMP_DUPLICATE_LIB_OK']='True'
record = 0
agent = Runner(game)
print("Now testing")
while agent.n_games != args.max_games:
state_old = agent.get_state()
dists = actorcritic(t(state_old).to(agent.device))[0]
actions = dists.sample().detach().cpu().data.numpy()
actions_clipped = np.clip(actions, -1, 1) #self.env.action_space.low.min(), env.action_space.high.max())
final_move = [0,0,0]
final_move[np.argmax(actions_clipped)] = 1
reward, done, score = game.playStep(final_move)
# final_move = agent.get_action(state_old, test_mode=True)
# reward, done, score = game.playStep(final_move)
if done:
agent.n_games += 1
game.gameReboot()
if score > record:
record = score
print('Game', agent.n_games, 'Score', score, 'Record:', record)
if __name__ == '__main__':
parser = argparse.ArgumentParser(description='RL Agent for Doodle Jump')
parser.add_argument("--macos", action="store_true", help="select model to train the agent")
parser.add_argument("--human", action="store_true", help="playing the game manually without agent")
parser.add_argument("--test", action="store_true", help="playing the game with a trained agent")
parser.add_argument("-d", "--difficulty", type=str, default="EASY", choices=["EASY", "MEDIUM", "HARD"], help="select difficulty of the game")
parser.add_argument("-m", "--model", type=str, default="a2c", choices=["a2c"], help="select model to train the agent")
parser.add_argument("-p", "--model_path", type=str, help="path to weights of an earlier trained model")
# parser.add_argument("-cp", "--critic_path", type=str, help="path to weights of an earlier trained model")
parser.add_argument("-alr", "--actor_lr", type=float, default=4e-4, help="set learning rate for training the model")
parser.add_argument("-clr", "--critic_lr", type=float, default=4e-3, help="set learning rate for training the model")
parser.add_argument("-g", "--gamma", type=float, default=0.9, help="set discount factor for q learning")
parser.add_argument("--max_memory", type=int, default=10000, help="Buffer memory size for long training")
parser.add_argument("--store_frames", action="store_true", help="store frames encountered during game play by agent")
parser.add_argument("--batch_size", type=int, default=1000, help="Batch size for long training")
parser.add_argument("--reward_type", type=int, default=1, choices=[1, 2, 3, 4], help="types of rewards formulation")
parser.add_argument("--exploration", type=int, default=40, help="number of games to explore")
parser.add_argument("--channels", type=int, default=1, help="set the image channels for preprocessing")
parser.add_argument("--height", type=int, default=80, help="set the image height post resize")
parser.add_argument("--width", type=int, default=80, help="set the image width post resize")
parser.add_argument("--server", action="store_true", help="when training on server add this flag")
parser.add_argument("--seed", type=int, default=42, help="change seed value for creating game randomness")
parser.add_argument("--max_games", type=int, default=1000, help="set the max number of games to be played by the agent")
args = parser.parse_args()
game = DoodleJump(difficulty=args.difficulty, server=args.server, reward_type=args.reward_type)
agent = Runner(game)
# env = gym.make("Pendulum-v0")
hyper_params = "_d_"+args.difficulty+"_m_"+args.model+"_alr_"+str(args.actor_lr)+"_clr_"+str(args.critic_lr)+"_g_"+str(args.gamma)+"_mem_"+str(args.max_memory)+"_batch_"+str(args.batch_size)
dstr = datetime.datetime.now().strftime("_dt-%Y-%m-%d-%H-%M-%S")
writer = SummaryWriter(log_dir="./model"+hyper_params+dstr)
arg_dict = vars(args)
writer.add_text('Model Parameters: ', str(arg_dict), 0)
# config
state = agent.get_state() #env.observation_space.shape[0]
n_actions = 3 #env.action_space.shape[0]
actorcritic = ActorCritic(state.shape[0], n_actions, activation=Mish).to(agent.device)
# critic = Critic(state.shape[0], activation=Mish).to(agent.device)
if (args.model_path) or args.test:
actorcritic.load_state_dict(torch.load(args.model_path))
# critic.load_state_dict(torch.load(args.critic_path))
if args.test:
test(game, args)
else:
learner = A2CLearner(actorcritic, agent.device, gamma=args.gamma, entropy_beta=0,
actor_lr=args.actor_lr, critic_lr=args.critic_lr, max_grad_norm=0.5, batch_size = args.batch_size)
# runner = Runner(env)
###########
steps_on_memory = 16
episodes = 500
episode_length = 200
total_steps = (episode_length*episodes)//steps_on_memory
# record = 0
while agent.n_games != args.max_games:
memory = agent.run(steps_on_memory)
learner.learn(memory, agent.steps, writer, discount_rewards=False)
writer.add_hparams(hparam_dict=vars(args),
metric_dict={'mean_reward': agent.mean_reward,
'high_score': agent.record,
'mean_score': agent.mean_score
})