config.py

embedding_dims = 50
MAX_FEATURES = 72039#184996
MAX_TEXT_LENGTH = 100
BATCH_SIZE = 32
EPOCHS = 2
VALIDATION_SPLIT = 0.1  
SPLIT=10000
SPLIT2=20000
CLASSES_LIST = ["toxic", "severe_toxic", "obscene", "threat", "insult", "identity_hate"]
path = './data/'

TRAIN_DATA_FILE=path+'train.csv'
TEST_DATA_FILE=path+'test.csv'
embedding_path=path+'crawl-300d-2M.vec.txt'
GLOVE_EMBEDDING_FILE=path+'glove.840B.300d.txt'

train_token_path='./train_corps.txt'
test_token_path='./test_corps.txt'

stop_words = {'the', 'a', 'an'}
stop_words.update(
    ['.', ',', '"', "'", '?', ':', ';', '(', ')', '[', ']', '{', '}','\'\'','``','...','-','%'])  # remove it if you need punctuation

badwords = {}

with open("data/badwords.txt") as fp:
    for line in fp:
        line = line.lower().strip()
        lines = line.split(',')
        if len(lines) == 1 and line not in badwords:
            badwords[line] = line
        elif len(lines) == 2 and lines[0] not in badwords:
            badwords[lines[0].strip()] = lines[1].strip().replace(" ", "_")

with open("data/stopwords.txt",encoding="utf-8") as fp:
    for line in fp:
        stop_words.add(line.strip())