transfromerForTasks/transformer_text_Classfication/hyperparams.py at master · fooSynaptic/transfromerForTasks · GitHub

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
# -*- coding: utf-8 -*-
#/usr/bin/python3


class rc_Hyperparams:
    trainset = './datasets/train_round_0.csv'
    testset = './datasets/test_data_r0.csv'

    trainfile = './preprocessed/train.csv'
    testfile = './preprocessed/test.csv'
    predictfile = './inference_QA.csv'

    batch_size = 64 # alias = N
    lr = 0.0001 # learning rate. In paper, learning rate is adjusted to the global step.
    logdir = 'rc_model_dir' # log directory

    # model
    q_maxlen = 50
    p_maxlen = 300
    ans_maxlen = 40
    min_cnt = 3 # words whose occurred less than min_cnt are encoded as <UNK>.
    hidden_units = 512 # alias = C
    num_blocks = 5 # number of encoder/decoder blocks
    num_epochs = 200
    num_heads = 8
    dropout_rate = 0.33
    sinusoid = False # If True, use sinusoid. If false, positional embedding.
    #n_class = 2
    dropout_keep_prob = 0.33
    reg_lambda = 0.1
    Passage_fuse = 'bi-rnn' # bi-rnn or Pooling
    use_dropout = True
    weight_decay = 0.1


class seq2seq_Hyperparams:
    '''Hyperparameters'''
    # data
    source_train = './datasets/zh-en/train.tags.zh-en.en'
    target_train = './datasets/zh-en/train.tags.zh-en.zh'
    source_test = './datasets/zh-en/IWSLT15.TED.tst2011.zh-en.en.xml'
    target_test = './datasets/zh-en/IWSLT15.TED.tst2011.zh-en.zh.xml'

    # training
    batch_size = 32 # alias = N
    lr = 0.0001 # learning rate. In paper, learning rate is adjusted to the global step.
    logdir = 'seq2seq_model_dir' # log directory

    # model
    maxlen = 100 # Maximum number of words in a sentence. alias = T.
                # Feel free to increase this if you are ambitious.
    min_cnt = 3 # words whose occurred less than min_cnt are encoded as <UNK>.
    hidden_units = 512 # alias = C
    num_blocks = 5 # number of encoder/decoder blocks
    num_epochs = 20
    num_heads = 8
    dropout_rate = 0.1
    sinusoid = False # If True, use sinusoid. If false, positional embedding.


class feature_Block_Hyperparams:
    '''Hyperparameters'''
    # data
    trainset = './datasets/cnews.train.txt'
    testset = './datasets/cnews.test.txt'
    tagging = {'时尚':0, '教育':1, '时政':2, '体育':3, '游戏':4, '家居':5, '科技':6, '房产':7, '财经':8, '娱乐':9}

    # training
    batch_size = 4 # alias = N
    lr = 0.0001 # learning rate. In paper, learning rate is adjusted to the global step.
    logdir = 'Block_model_dir' # log directory

    # model
    maxlen = 500 # Maximum number of words in a sentence. alias = T.
                # Feel free to increase this if you are ambitious.
    min_cnt = 3 # words whose occurred less than min_cnt are encoded as <UNK>.
    hidden_units = 512 # alias = C
    num_blocks = 5 # number of encoder/decoder blocks
    num_epochs = 20
    num_heads = 8
    dropout_rate = 0.1
    sinusoid = False # If True, use sinusoid. If false, positional embedding.
    n_class = 10


class infersent_Block_Hyperparams:
    '''Hyperparameters'''
    # data
    trainset = './opensrc_dta/train.csv'
    testset = './opensrc_dta/test.csv'


    # training
    relations = {'entailment': '0', 'contradiction': '1', 'neutral': '2'}

    batch_size = 64 # alias = N
    lr = 0.0001 # learning rate. In paper, learning rate is adjusted to the global step.
    logdir = 'infersent_model_dir' # log directory

    # model
    maxlen = 24 # Maximum number of words in a sentence. alias = T.
                # Feel free to increase this if you are ambitious.
    min_cnt = 3 # words whose occurred less than min_cnt are encoded as <UNK>.
    hidden_units = 512 # alias = C
    num_blocks = 5 # number of encoder/decoder blocks
    num_epochs = 20
    num_heads = 8
    dropout_rate = 0.1
    sinusoid = False # If True, use sinusoid. If false, positional embedding.
    #n_class = 2
    dropout_keep_prob = 0.55
    reg_lambda = 0.1