-
Notifications
You must be signed in to change notification settings - Fork 0
/
utils.py
154 lines (122 loc) · 6.15 KB
/
utils.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
from player.player import Player
from environments.simulator import Atari
import numpy as np
import datetime
import os
import json
import csv
import imageio
from skimage.transform import resize
MAX_EPISODE_LENGTH= 18000
NO_OP_STEPS= 10
MAX_EPISODES= 3000
AGENT_HISTORY_LENGTH= 4
UPDATE_FREQ= 4
NETW_UPDATE_FREQ= 10000
REPLAY_MEMORY_START_SIZE = 50000
DISCOUNT_FACTOR= 0.99
MEMORY_SIZE = 1000000
BS= 32
LEARNING_RATE= 0.0001
PUNISH= 0.0
INI_EPSILON= 1.0
END_EPSILON= 0.1
MIN_OBSERVE_EPISODE= 200
GAME_ENV= "BreakoutDeterministic-v4"
REWARD_EXTRAPOLATION_EXPONENT = -5.0
frame_height = 84
frame_width = 84
LINEAR_EXPLORATION_EXPONENT = False
USE_DOUBLE_MODEL = True
class HandleResults:
folder_to_use = ''
settings_file_name = 'settings.jsn'
time = datetime.datetime.now()
def __init__(self, game_env, out_folder):
# GAME_ENV = settings['GAME_ENV']
d = datetime.datetime.now().strftime("%y_%m_%d_%H_%M_%S")
self.folder_to_use = ''.join([out_folder, game_env, '/results_', d, '/'])
if not os.path.exists(self.folder_to_use):
os.makedirs(self.folder_to_use)
self.results_file_name = ''.join([self.folder_to_use, 'results_', game_env, '_', d, '.csv'])
def save_settings(self, settings, player= None):
settings_dict = settings
with open(''.join([self.folder_to_use, 'settings.jsn']), 'wt') as outfile:
json.dump(settings_dict, outfile, indent=4)
if player is not None:
player.save_player_learner(self.folder_to_use)
def load_settings_default(self, GAME_ENV):
settings_dict = {}
file_name = './default_settings.jsn' # default_settings.jsn i in the root
with open(file_name, 'rt') as json_file:
settings_dict = json.load(json_file)
return self.load_settings_dictionary(GAME_ENV, settings_dict)
def load_settings_dictionary(self, GAME_ENV, settings_dict):
settings_dict['GAME_ENV'] = GAME_ENV
game_env = Atari(settings_dict['GAME_ENV'], settings_dict['frame_height'], settings_dict['frame_width'],
agent_history_length=settings_dict['AGENT_HISTORY_LENGTH'],
no_op_steps=settings_dict['NO_OP_STEPS'])
player = self.build_player(settings_dict, game_env)
return player, game_env, settings_dict['MAX_EPISODE_LENGTH'], settings_dict['MAX_EPISODES'], settings_dict
def load_default_settings_constants(self, GAME_ENV):
settings_dict = {}
settings_dict['GAME_ENV'] = GAME_ENV
settings_dict['AGENT_HISTORY_LENGTH'] = AGENT_HISTORY_LENGTH
settings_dict['MEMORY_SIZE'] = MEMORY_SIZE
settings_dict['BS'] = BS
settings_dict['LEARNING_RATE'] = LEARNING_RATE
settings_dict['INI_EPSILON'] = INI_EPSILON
settings_dict['END_EPSILON'] = END_EPSILON
settings_dict['MIN_OBSERVE_EPISODE'] = MIN_OBSERVE_EPISODE
settings_dict['NETW_UPDATE_FREQ'] = NETW_UPDATE_FREQ
settings_dict['UPDATE_FREQ'] = UPDATE_FREQ
settings_dict['DISCOUNT_FACTOR'] = DISCOUNT_FACTOR
settings_dict['REPLAY_MEMORY_START_SIZE'] = REPLAY_MEMORY_START_SIZE
settings_dict['PUNISH'] = PUNISH
settings_dict['REWARD_EXTRAPOLATION_EXPONENT'] = REWARD_EXTRAPOLATION_EXPONENT
settings_dict['LINEAR_EXPLORATION_EXPONENT'] = LINEAR_EXPLORATION_EXPONENT
settings_dict['USE_DOUBLE_MODEL'] = USE_DOUBLE_MODEL
settings_dict['frame_height'] = frame_height
settings_dict['frame_width'] = frame_width
settings_dict['NO_OP_STEPS'] = NO_OP_STEPS
settings_dict['MAX_EPISODE_LENGTH'] = MAX_EPISODE_LENGTH
settings_dict['MAX_EPISODES'] = MAX_EPISODES
return self.load_settings_dictionary(GAME_ENV, settings_dict)
def load_settings_folder(self, folder, load_model):
settings_dict = {}
with open(''.join([folder, self.settings_file_name]), 'rt') as json_file:
settings_dict = json.load(json_file)
player, game_env, _, _, _ = self.load_settings_dictionary(GAME_ENV, settings_dict)
if load_model:
player.load_player_learner(folder)
return player, game_env, settings_dict['MAX_EPISODE_LENGTH'], settings_dict['MAX_EPISODES'], settings_dict
def save_res(self, res_dict):
if not os.path.isfile(self.results_file_name):
with open(self.results_file_name, mode='w', newline='') as file:
writer = csv.writer(file)
headings = list(res_dict.keys())
writer.writerow(headings)
file.close()
with open(self.results_file_name, mode='a', newline='') as file:
writer = csv.writer(file)
values = list(res_dict.values())
writer.writerow(values)
file.close()
print(res_dict)
def generate_gif(self, episode, frames_for_gif, reward, string_to_add):
path = self.folder_to_use
for idx, frame_idx in enumerate(frames_for_gif):
frames_for_gif[idx] = resize(frame_idx, (420, 320, 3),
preserve_range=True, order=0).astype(np.uint8)
imageio.mimsave(f'{path}{"ATARI_episode_{0}_reward_{1}_{2}.gif".format(episode, reward, string_to_add)}',
frames_for_gif, duration=1 / 30)
def build_player(self, settings_dict, game_env):
player = Player(game_env, settings_dict['AGENT_HISTORY_LENGTH'], settings_dict['MEMORY_SIZE'],
settings_dict['BS'],
settings_dict['LEARNING_RATE'], settings_dict['INI_EPSILON'], settings_dict['END_EPSILON'],
settings_dict['MIN_OBSERVE_EPISODE'], settings_dict['NETW_UPDATE_FREQ'],
settings_dict['UPDATE_FREQ'], settings_dict['DISCOUNT_FACTOR'],
settings_dict['REPLAY_MEMORY_START_SIZE'], settings_dict['PUNISH'],
settings_dict['REWARD_EXTRAPOLATION_EXPONENT'], settings_dict['LINEAR_EXPLORATION_EXPONENT'],
settings_dict['USE_DOUBLE_MODEL'])
return player