-
Notifications
You must be signed in to change notification settings - Fork 3
/
Config.py
62 lines (53 loc) · 1.84 KB
/
Config.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
"""
find which config to use and select
"""
import json
def Config(tianchivect, flag=None):
if flag == None:
return config_init(tianchivect)
elif flag == "haha":
return haha()
else:
raise ValueError("Invalid model: %s", flag)
class config_init(object):
epoch_num = 30000
max_grad_norm = 1
init_scale = 0.05
learning_rate = 0.001 # 0.2 0.01
adagrad_eps = 1e-5
DIM = 128 # Model dimensionality
OUTPUT_DIM = 100
D_hidden_layer = 1 # 10
disc_iter = 5
gen_iter = 1
# category_num = 71
BATCH_SIZE = 30 # Batch size # set it to 2 and the normalization could be in easy
CRITIC_ITERS = 5 # For WGAN and WGAN-GP, number of critic iters per gen iter
LAMBDA = 10 # Gradient penalty lambda hyperparameter
delta = 0.01 # the diversity
ITERS = 1000 # How many generator iterations to train for
GIMG = 4096
ratio_super = (0.5, 0.5) # supervise : unsupervise
gpuid = 1
def __init__(self, tianchivect):
# self.vocab_size = len(reader.vocab.words) # 10000
print "config~"
print "learning_rate", self.learning_rate
print "max_grad_norm", self.max_grad_norm
print "disc_iter", self.disc_iter
print "gen_iter", self.gen_iter
print "DIM", self.DIM
print "delta", self.delta
print "ratio_super", self.ratio_super
with open(tianchivect.filepath + "cate_2match.json", 'r') as f:
cate_2match = json.load(f)
self.cat_match = len(cate_2match)
self.catmatch_id = cate_2match
self.cate2id = tianchivect.cate2id
self.category_num = len(tianchivect.cate2id)
print tianchivect.filepath
print "category_num", self.category_num
print "cat_match_num", self.cat_match
class haha(object):
def __init__(self, loader):
print "haha~"