-
Notifications
You must be signed in to change notification settings - Fork 2
/
config_args.py
148 lines (121 loc) · 7.27 KB
/
config_args.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
import os
from pdb import set_trace as stop
def get_args(parser,eval=False):
"""Gets the arguments for the parser, sets the arguments and tasks, and sets the environment.
The method initializes the training and validation datasets, saves the models, and creates directories for the models.
:param parser: Contains the parser variable, has parse_args
:param eval: Always set to false
"""
parser.add_argument("-tr", "--train_dataset", type=str, default=None, help="train dataset for train bert")
parser.add_argument("-va", "--valid_dataset", type=str, default=None, help="test set for evaluate train set")
parser.add_argument("-te", "--test_dataset", type=str, default=None, help="test set for evaluate train set")
parser.add_argument("-v", "--vocab_path", required=True, type=str, help="built vocab model path with bert-vocab")
parser.add_argument("-o", "--output_path", required=True, type=str, help="output/bert.model")
parser.add_argument("-d", "--data_root", type=str, default='./dataset/data/', help="label file for ss task")
parser.add_argument("--activation", type=str, default='gelu',choices=['relu', 'gelu'])
parser.add_argument("--task", type=str, default='secondary', help="")
parser.add_argument('--tasks', nargs='+',default='')
parser.add_argument("-hs", "--hidden", type=int, default=768, help="hidden size of transformer model")
parser.add_argument("-l", "--layers", type=int, default=12, help="number of layers")
parser.add_argument("-a", "--attn_heads", type=int, default=8, help="number of attention heads")
parser.add_argument("-do", "--dropout", type=float, default=0.1, help="dropout")
parser.add_argument("-s", "--seq_len", type=int, default=1024, help="maximum sequence len")
parser.add_argument("-s2", "--seq_len2", type=int, default=-1, help="maximum sequence len for testing")
parser.add_argument("-b", "--batch_size", type=int, default=16, help="number of batch_size")
parser.add_argument("--max_batches", type=int, default=-1, help="")
parser.add_argument("--test_batch_size", type=int, default=-1, help="number of batch_size")
parser.add_argument("-e", "--epochs", type=int, default=100, help="number of epochs")
parser.add_argument("-w", "--num_workers", type=int, default=0, help="dataloader worker size")
parser.add_argument("--with_cuda", type=bool, default=True, help="training with CUDA: true, or false")
parser.add_argument("--log_freq", type=int, default=1, help="printing loss every n iter: setting n")
parser.add_argument("--corpus_lines", type=int, default=None, help="total number of lines in corpus")
parser.add_argument("--cuda_devices", type=int, nargs='+', default=None, help="CUDA device ids")
parser.add_argument("--on_memory", type=bool, default=True, help="Loading on memory: true or false")
parser.add_argument("--overwrite", action='store_true')
parser.add_argument("--dont_overwrite", action='store_true')
parser.add_argument("--freeze_bert", action='store_true')
parser.add_argument("--esm", action='store_true')
parser.add_argument("--reset_weights", action='store_true')
parser.add_argument("--lr", type=float, default=2e-4, help="learning rate of adam")
parser.add_argument("--adam_weight_decay", type=float, default=0, help="weight_decay of adam")
parser.add_argument("--optimizer", type=str, default='adam',choices=['sgd', 'adam'])
parser.add_argument("--adam_beta1", type=float, default=0.9, help="adam first beta value")
parser.add_argument("--adam_beta2", type=float, default=0.999, help="adam first beta value")
parser.add_argument("--warmup_steps", type=int, default=1000, help="warmup_steps")
parser.add_argument("--name", type=str, default='', help="")
parser.add_argument("-save_root", "--save_root", type=str, default='./results/', help="")
parser.add_argument('--debug', action='store_true')
parser.add_argument('--pretrain', action='store_true')
parser.add_argument('--test_lm', action='store_true')
parser.add_argument("--grad_ac_steps", type=int, default=1, help="")
parser.add_argument("--fold", type=int, default=1, help="")
parser.add_argument('--use_evo', action='store_true')
parser.add_argument("--saved_bert", type=str, default='', help="")
parser.add_argument("--saved_model", type=str, default='', help="")
parser.add_argument("--emb_type", type=str, default='conv',choices=['lookup','conv','continuous','both','pair'])
args = parser.parse_args()
args.save_best = True
if args.debug:
# args.train_dataset = os.path.join(args.data_root,args.test_dataset)
args.train_dataset = os.path.join(args.data_root,args.valid_dataset)
args.valid_dataset = None
args.test_dataset = None
args.batch_size = 4
args.cuda_devices = [0]
else:
args.train_dataset = os.path.join(args.data_root,args.train_dataset) if args.train_dataset is not None else None
args.valid_dataset = os.path.join(args.data_root,args.valid_dataset) if args.valid_dataset is not None else None
args.test_dataset = os.path.join(args.data_root,args.test_dataset) if args.test_dataset is not None else None
args.vocab_path = os.path.join(args.data_root,args.vocab_path)
if args.seq_len2 == -1:
args.seq_len2 = args.seq_len
args.model_name = os.path.join(args.save_root,args.task)
args.model_name += '.bert'
args.model_name += '.bsz_'+str(int(args.batch_size*args.grad_ac_steps))
args.model_name += '.layers_'+str(args.layers)
args.model_name += '.size_'+str(args.hidden)
args.model_name += '.heads_'+str(args.attn_heads)
args.model_name += '.drop_'+("%.2f" % args.dropout).split('.')[1]
args.model_name += '.lr_'+str(args.lr)#.split('.')[1]
if args.test_batch_size != -1:
args.batch_size = args.test_batch_size
if args.use_evo:
args.model_name += '.use_evo'
args.emb_type = 'both' # fix
args.model_name += '.emb_'+args.emb_type
if args.debug:
args.model_name += '.debug'
if args.saved_bert != '':
args.model_name += '.saved_bert'
if args.reset_weights:
args.model_name += '.reset_weights'
if args.saved_model != '':
args.model_name += '.saved_model'
if args.esm:
args.model_name += '.esm'
if args.freeze_bert:
args.model_name += '.freeze_bert'
if args.train_dataset:
if 'h1n1' in args.train_dataset:
args.model_name += '.h1n1'
elif 'ebola' in args.train_dataset:
args.model_name += '.ebola'
elif 'barman' in args.train_dataset:
args.model_name += '.barman'
elif 'denovo' in args.train_dataset.lower():
args.model_name += '.denovo'
if args.name != '':
args.model_name += '.'+args.name
if os.path.isdir(args.model_name) and args.dont_overwrite:
exit(0)
if os.path.isdir(args.model_name) and (not args.overwrite) and (not args.debug):
print(args.model_name)
overwrite_status = input('Already Exists. Overwrite?: ')
if overwrite_status == 'rm':
os.system('rm -rf '+args.model_name)
elif not 'y' in overwrite_status:
exit(0)
if not os.path.exists(args.model_name):
os.makedirs(args.model_name)
args.evo_size = 30
return args