def preprocess_data(vocab_size, batch_size, num_workers=0, max_input_len=400,
max_target_len=100):
p = Preprocessor(chosen_dataset)
print('preprocessing started')
train_set, test_set, validation_set = p.create_data_loaders(vocab_size, batch_size, num_workers=num_workers,
max_input_len=max_input_len, max_target_len=max_target_len)
print('preprocessing finished')
return p, train_set, test_set, validation_set
