This repository has been archived by the owner on Oct 31, 2023. It is now read-only.
-
Notifications
You must be signed in to change notification settings - Fork 7
/
main.py
123 lines (111 loc) · 4.84 KB
/
main.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
import argparse
from eos import EOS
def parse_arguments():
parser = argparse.ArgumentParser()
parser.add_argument("mode", choices=["train", "test", "tag", "extract"])
parser.add_argument("--training-file",
help="Defines training data set")
parser.add_argument("--test-file",
help="Defines test data set")
parser.add_argument("--input-file",
help="Defines input file to be tagged")
parser.add_argument("--epochs", default=5,
help="Defines number of training epochs")
parser.add_argument(
"--architecture",
default="cnn",
help="Neural network architectures, supported: cnn, lstm, bi-lstm, gru, bi-gru, mlp")
parser.add_argument("--window-size", default=5,
help="Defines number of window size (char-ngram)")
parser.add_argument("--batch-size", default=32,
help="Defines number of batch_size")
parser.add_argument("--dropout", default=0.2,
help="Defines number dropout")
parser.add_argument(
"--min-freq",
default=100,
help="Defines the min. freq. a char must appear in data")
parser.add_argument("--max-features", default=200,
help="Defines number of features for Embeddings layer")
parser.add_argument("--embedding-size", default=128,
help="Defines Embeddings size")
parser.add_argument("--kernel-size", default=8,
help="Defines Kernel size of CNN")
parser.add_argument("--filters", default=6,
help="Defines number of filters of CNN")
parser.add_argument("--pool-size", default=8,
help="Defines pool size of CNN")
parser.add_argument("--hidden-dims", default=250,
help="Defines number of hidden dims")
parser.add_argument("--strides", default=1,
help="Defines numer of strides for CNN")
parser.add_argument("--lstm_gru_size", default=256,
help="Defines size of LSTM/GRU layer")
parser.add_argument("--mlp-dense", default=6,
help="Defines number of dense layers for mlp")
parser.add_argument("--mlp-dense-units", default=16,
help="Defines number of dense units for mlp")
parser.add_argument("--model-filename", default='best_model.hdf5',
help="Defines model filename")
parser.add_argument("--vocab-filename", default='vocab.dump',
help="Defines vocab filename")
parser.add_argument("--eos-marker", default='</eos>',
help="Defines end-of-sentence marker used for tagging")
args = parser.parse_args()
nn_eos = EOS()
if args.mode == "train":
if not args.training_file:
print("Training data file name is missing!")
parser.print_help()
exit(1)
nn_eos.train(args.training_file,
str(args.architecture),
int(args.window_size),
int(args.epochs),
int(args.batch_size),
float(args.dropout),
int(args.min_freq),
int(args.max_features),
int(args.embedding_size),
int(args.lstm_gru_size),
int(args.mlp_dense),
int(args.mlp_dense_units),
int(args.kernel_size),
int(args.filters),
int(args.pool_size),
int(args.hidden_dims),
int(args.strides),
args.model_filename,
args.vocab_filename)
elif args.mode == "test":
if not args.test_file:
print("Test data file name is missing!")
parser.print_help()
exit(1)
nn_eos.test(args.test_file,
args.model_filename,
args.vocab_filename,
int(args.window_size),
int(args.batch_size))
elif args.mode == "tag":
if not args.input_file:
print("Input file name is missing!")
parser.print_help()
exit(1)
nn_eos.tag(args.input_file,
args.model_filename,
args.vocab_filename,
int(args.window_size),
int(args.batch_size),
args.eos_marker)
elif args.mode == "extract":
if not args.input_file:
print("Input file name is missing!")
parser.print_help()
exit(1)
nn_eos.extract(
args.input_file, int(
args.window_size), int(
args.min_freq))
if __name__ == '__main__':
parse_arguments()