-
Notifications
You must be signed in to change notification settings - Fork 0
/
settings.py
101 lines (66 loc) · 2.36 KB
/
settings.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
from multiprocessing import cpu_count
class Settings:
###########################################################################
# Environment settings
ENV = "Pendulum-v0"
LOAD = True
DISPLAY = True
GUI = True
TRAINING_EPS = 1000
MAX_EPISODE_STEPS = 15000
FRAME_SKIP = 3
EP_ELONGATION = 10
###########################################################################
# Network settings
# CONV_LAYERS = [
# {'filters': 32, 'kernel_size': [8, 8], 'strides': [4, 4]},
# {'filters': 64, 'kernel_size': [4, 4], 'strides': [2, 2]},
# {'filters': 64, 'kernel_size': [3, 3], 'strides': [1, 1]}
# ]
HIDDEN_ACTOR_LAYERS = [8, 8, 8]
HIDDEN_CRITIC_LAYERS = [8, 8, 8]
NB_ATOMS = 51
ACTOR_LEARNING_RATE = 5e-4
CRITIC_LEARNING_RATE = 5e-4
###########################################################################
# Algorithm hyper-parameters
NB_ACTORS = 64 # cpu_count() - 2
DISCOUNT = 0.99
N_STEP_RETURN = 1
DISCOUNT_N = DISCOUNT ** N_STEP_RETURN
MIN_Q = -2000
MAX_Q = 0
BUFFER_SIZE = 100000
BATCH_SIZE = 64
UPDATE_TARGET_FREQ = 1
UPDATE_TARGET_RATE = 0.05
UPDATE_ACTORS_FREQ = 1
###########################################################################
# Exploration settings
NOISE_SCALE = 0.3
NOISE_DECAY = 0.99
###########################################################################
# Features frequencies
EP_REWARD_FREQ = 50
PLOT_FREQ = 100
RENDER_FREQ = 1000
GIF_FREQ = 2000
SAVE_FREQ = 1000
PERF_FREQ = 100
###########################################################################
# Save settings
RESULTS_PATH = 'results/'
MODEL_PATH = 'model/'
GIF_PATH = 'results/gif/'
MAX_NB_GIF = 5
###########################################################################
import gym
setting_env = gym.make(ENV)
if 'CONV_LAYERS' in locals():
STATE_SIZE = [84, 84, 4]
else:
STATE_SIZE = list(setting_env.observation_space.shape)
ACTION_SIZE = setting_env.action_space.shape[0]
LOW_BOUND = setting_env.action_space.low
HIGH_BOUND = setting_env.action_space.high
del setting_env