settings.py

import tensorflow as tf
import numpy as np

flags = tf.app.flags
FLAGS = flags.FLAGS

flags.DEFINE_integer('hidden3', 64, 'Number of units in hidden layer 3.')
flags.DEFINE_float('learning_rate', 3e-5, 'Initial learning rate.')
flags.DEFINE_integer('hidden1', 4096, 'Number of units in hidden layer 1.')
flags.DEFINE_integer('hidden2', 4096, 'Number of units in hidden layer 2.')
flags.DEFINE_integer('input_view', 0, 'View No. informative view, ACM:0, DBLP:1')
flags.DEFINE_float('weight_decay', 0.0001, 'Weight for L2 loss on embedding matrix.')
flags.DEFINE_float('fea_decay', 0.5, 'feature decay.')
flags.DEFINE_float('weight_R', 0.001, 'Weight for R loss on embedding matrix.')
flags.DEFINE_float('dropout', 0., 'Dropout rate (1 - keep probability).')
flags.DEFINE_float('attn_drop', 0., 'Dropout rate (1 - keep probability).')
flags.DEFINE_float('ffd_drop', 0., 'Dropout rate (1 - keep probability).')
flags.DEFINE_integer('features', 1, 'Whether to use features (1) or not (0).')
flags.DEFINE_integer('seed', 50, 'seed for fixing the results.')
flags.DEFINE_integer('pre_iterations', 1, 'number of pre_iterations.')
flags.DEFINE_integer('fin_iterations', 100, 'number of fin_iterations.')
flags.DEFINE_integer('n_clusters', 3, 'predict label early stop.')
flags.DEFINE_float('kl_decay', 0.1, 'kl loss decay.')

# infor: number of clusters
infor = {'ACM': 3, 'DBLP': 4}


def get_settings(dataname, model, task):
    global re
    if dataname != 'ACM' and dataname != 'DBLP':
        print('error: wrong data set name')
    if model != 'Main':
        print('error: wrong model name')
    if task != 'clustering':
        print('error: wrong task name')

    if task == 'clustering':
        pre_iterations = FLAGS.pre_iterations
        fin_iterations = FLAGS.fin_iterations
        clustering_num = infor[dataname]
        re = {'data_name': dataname, 'pre_iterations': pre_iterations, 'fin_iterations': fin_iterations,
              'clustering_num': clustering_num, 'model': model}
    return re