-
Notifications
You must be signed in to change notification settings - Fork 0
/
helper.py
165 lines (124 loc) · 5.48 KB
/
helper.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
"""
helper.py
"""
__author__ = "[email protected]"
import time
import os
import json
import sys
import argparse
from collections import OrderedDict
import numpy as np
def experiment_load():
parser = argparse.ArgumentParser()
parser.add_argument("CLUSTER")
parser.add_argument("EXPERIMENT")
args = parser.parse_args()
# let experiment type (function) come from commandline arg
with open(args.EXPERIMENT) as jconfig:
DDPG_config = json.load(jconfig)
DDPG_config['CLUSTER'] = args.CLUSTER
DDPG_config['EXPERIMENT'] = args.EXPERIMENT.lower().split('.')[0]
if DDPG_config['CLUSTER'] == 'local':
import experiment.local
runwrapper = experiment.local.runwrapper
DDPG_config['EXPERIMENT'] = setup_exp(DDPG_config['EXPERIMENT'])
return DDPG_config, runwrapper
def setup_exp(experiment=''):
folder = 'runs/'
os.makedirs(folder, exist_ok=True)
folder += experiment + '/'
os.makedirs(folder, exist_ok=True)
return folder
def setup_run(DDPG_config):
folder = DDPG_config['EXPERIMENT']
epoch = 't%.6f/' % time.time()
folder += epoch.replace('.', '')
os.makedirs(folder, exist_ok=True)
with open(folder + 'folder.ini', 'w') as ifile:
ifile.write('[General]\n')
ifile.write('**.folderName = "' + folder + '"\n')
with open(folder + 'DDPG.json', 'w') as jconfig:
json.dump(OrderedDict(sorted(DDPG_config.items(), key=lambda t: t[0])), jconfig, indent=4)
# with open(folder + 'Routing.txt', 'w') as rfile:
# rfile.write(DDPG_config['U_ROUTING'] + '\n')
if DDPG_config['TRAFFIC'].startswith('STAT:'):
with open(folder + 'Traffic.txt', 'w') as rfile:
rfile.write(DDPG_config['TRAFFIC'].split('STAT:')[-1] + '\n')
return folder
def setup_brute(DDPG_config):
folder = 'runs/brute'
epoch = 't%.6f/' % time.time()
folder += epoch.replace('.', '')
folder += '/'
os.makedirs(folder, exist_ok=True)
with open(folder + 'folder.ini', 'w') as ifile:
ifile.write('[General]\n')
ifile.write('**.folderName = "' + folder + '"\n')
# with open(folder + 'Routing.txt', 'w') as rfile:
# rfile.write(DDPG_config['U_ROUTING'] + '\n')
if DDPG_config['TRAFFIC'].startswith('STAT:'):
with open(folder + 'Traffic.txt', 'w') as rfile:
rfile.write(DDPG_config['TRAFFIC'].split('STAT:')[-1] + '\n')
with open(folder + 'DDPG.json', 'w') as jconfig:
json.dump(OrderedDict(sorted(DDPG_config.items(), key=lambda t: t[0])), jconfig, indent=4)
return folder
def parser():
parser = argparse.ArgumentParser()
parser.add_argument("CLUSTER")
parser.add_argument("EXPERIMENT")
parser.add_argument("--RSEED", type=int, action="store", default=None)
parser.add_argument("--PRINT", action="store_true")
parser.add_argument("--ACTIVE_NODES", type=int, action="store", required=True)
parser.add_argument("--MU", type=float, action="store", required=True)
parser.add_argument("--THETA", type=float, action="store", required=True)
parser.add_argument("--SIGMA", type=float, action="store", required=True)
parser.add_argument("--BUFFER_SIZE", type=int, action="store", required=True)
parser.add_argument("--BATCH_SIZE", type=int, action="store", required=True)
parser.add_argument("--GAMMA", type=float, action="store", required=True)
parser.add_argument("--TAU", type=float, action="store", required=True)
parser.add_argument("--LRA", type=float, action="store", required=True)
parser.add_argument("--LRC", type=float, action="store", required=True)
parser.add_argument("--EXPLORE", type=float, action="store", required=True)
parser.add_argument("--EPISODE_COUNT", type=int, action="store", required=True)
parser.add_argument("--MAX_STEPS", type=int, action="store", required=True)
parser.add_argument("--HACTI", action="store", required=True)
parser.add_argument("--HIDDEN1_UNITS", type=int, action="store", required=True)
parser.add_argument("--HIDDEN2_UNITS", type=int, action="store", required=True)
parser.add_argument("--TRAFFIC", action="store", required=True)
parser.add_argument("--STATUM", action="store", required=True)
parser.add_argument("--PRAEMIUM", action="store", required=True)
parser.add_argument("--ACTUM", action="store", required=True)
parser.add_argument("--MAX_DELTA", type=float, action="store", default=None)
parser.add_argument("--BN", action="store", default=None)
parser.add_argument("--U_ROUTING", action="store", default=None)
parser.add_argument("--ROUTING", action="store", required=True)
parser.add_argument("--ENV", action="store", required=True)
args = parser.parse_args()
DDPG_config = vars(args)
return DDPG_config
def pretty(f):
try:
float(f)
return str.format('{0:.3f}', f).rstrip('0').rstrip('.')
except:
return str(f)
def scale(array):
mean = array.mean()
std = array.std()
if std == 0:
std = 1
return np.asarray((array - mean)/std)
def softmax(x):
return np.exp(x) / np.sum(np.exp(x), axis=0)
def selu(x):
from keras.activations import elu
"""Scaled Exponential Linear Unit. (Klambauer et al., 2017)
# Arguments
x: A tensor or variable to compute the activation function for.
# References
- [Self-Normalizing Neural Networks](https://arxiv.org/abs/1706.02515)
"""
alpha = 1.6732632423543772848170429916717
scale = 1.0507009873554804934193349852946
return scale * elu(x, alpha)