config.py

import argparse
import inspect
import torch


class Config:
    word2vec_file = 'embedding/numberbatch-en.txt'
    emotion_file = 'embedding/counter_fitted_vector_space-0.txt'
    train_file = 'dataset/Twitter/train.json'
    valid_file = 'dataset/Twitter/valid.json'
    test_file = 'dataset/Twitter/test.json'
    study_file = 'dataset/visiont/test_10.json'
    saved_model = 'model/best_model.pt'
    device = torch.device("cuda:0")
    train_epochs = 200
    batch_size = 32
    learning_rate = 0.001  #or 0.005
    learning_rate_decay = 0.99
    l2_regularization = 0.001
    content_count = 1  # max count of content
    content_length = 32  # max count of news content words
    comment_count = 10  # max count of user comments
    review_length = 32  # max count of comment words
    lowest_review_count = 3  # Minimum number of comments for users to keep
    PAD_WORD = '<UNK>'
    require_improvment = 1000
    #class_list = ['fake', 'real']   # weibo
    class_list = ['false', 'true']  # twitter