classifier_multi_label_seq2seq_attention/hyperparameters.py at main · hellonlp/classifier_multi_label_seq2seq_attention · 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
# -*- coding: utf-8 -*-
"""
Created on Mon Nov 12 14:23:12 2018

@author: cm
"""

import os
import sys

pwd = os.path.dirname(os.path.abspath(__file__))
sys.path.append(pwd)
from classifier_multi_label_seq2seq_attention.utils import load_vocabulary


class Hyperparamters:
    # Train parameters
    num_train_epochs = 20
    print_step = 100
    batch_size = 4  # 128
    summary_step = 10
    num_saved_per_epoch = 3
    max_to_keep = 100
    logdir = 'logdir/CML_Seq2Seq'
    file_save_model = 'model/CML_Seq2Seq'
    file_load_model = 'model/load_01'

    # Train/Test data
    data_dir = os.path.join(pwd, 'data')
    train_data = 'train.csv'
    test_data = 'test.csv'

    # Load vocabulcary dict
    dict_id2label, dict_label2id = load_vocabulary(os.path.join(pwd, 'data', 'vocabulary_label.txt'))
    label_vocabulary = list(dict_id2label.values())

    # Optimization parameters
    warmup_proportion = 0.1
    use_tpu = None
    do_lower_case = True
    learning_rate = 5e-5

    # BiLSTM parameters
    num_layer_lstm_encode = 3
    lstm_hidden_size = 768
    decoder_embedding_size = 768

    # Beam search
    is_beam_search = True
    beam_size = 5
    max_length = 5

    # Sequence and Label
    sequence_length = 60
    num_labels = len(list(dict_id2label))

    # ALBERT
    model = 'albert_small_zh_google'
    bert_path = os.path.join(pwd, model)
    vocab_file = os.path.join(pwd, model, 'vocab_chinese.txt')
    init_checkpoint = os.path.join(pwd, model, 'albert_model.ckpt')
    saved_model_path = os.path.join(pwd, 'model')