How to use the fairseq.utils.import_user_module function in fairseq

def main(args):
    assert args.path is not None, '--path required for generation!'
    assert not args.sampling or args.nbest == args.beam, \
        '--sampling requires --nbest to be equal to --beam'

    utils.import_user_module(args)

    if args.max_tokens is None and args.max_sentences is None:
        args.max_tokens = 12000
    print(args)

    use_cuda = torch.cuda.is_available() and not args.cpu

    # Load dataset split
    task = tasks.setup_task(args)
    task.load_dataset(args.gen_subset)

    # Load ensemble
    print('| loading model(s) from {}'.format(args.path))
    model_paths = args.path.split(':')
    models, _model_args = checkpoint_utils.load_model_ensemble(
        model_paths,

def main(args, init_distributed=False):
    utils.import_user_module(args)

    assert args.max_tokens is not None or args.max_sentences is not None, \
        'Must specify batch size either with --max-tokens or --max-sentences'

    # Initialize CUDA and distributed training
    if torch.cuda.is_available() and not args.cpu:
        torch.cuda.set_device(args.device_id)
    np.random.seed(args.seed)
    torch.manual_seed(args.seed)
    if init_distributed:
        args.distributed_rank = distributed_utils.distributed_init(args)

    if distributed_utils.is_master(args):
        checkpoint_utils.verify_checkpoint_directory(args.save_dir)

    # Print args

def main(args, init_distributed=False):
    utils.import_user_module(args)

    assert args.max_tokens is not None or args.max_sentences is not None, \
        'Must specify batch size either with --max-tokens or --max-sentences'

    # Initialize CUDA and distributed training
    if torch.cuda.is_available() and not args.cpu:
        torch.cuda.set_device(args.device_id)
    np.random.seed(args.seed)
    torch.manual_seed(args.seed)
    if init_distributed:
        args.distributed_rank = distributed_utils.distributed_init(args)

    if distributed_utils.is_master(args):
        checkpoint_utils.verify_checkpoint_directory(args.save_dir)

    # Print args

def main(args):
    utils.import_user_module(args)

    if args.buffer_size < 1:
        args.buffer_size = 1
    if args.max_tokens is None and args.max_sentences is None:
        args.max_sentences = 1

    assert not args.sampling or args.nbest == args.beam, \
        '--sampling requires --nbest to be equal to --beam'
    assert not args.max_sentences or args.max_sentences <= args.buffer_size, \
        '--max-sentences/--batch-size cannot be larger than --buffer-size'

    print(args)

    use_cuda = torch.cuda.is_available() and not args.cpu

    # Setup task, e.g., translation

def main(args):
    import_user_module(args)

    if args.buffer_size < 1:
        args.buffer_size = 1
    if args.max_tokens is None and args.max_sentences is None:
        args.max_sentences = 1

    assert not args.sampling or args.nbest == args.beam, \
        '--sampling requires --nbest to be equal to --beam'
    assert not args.max_sentences or args.max_sentences <= args.buffer_size, \
        '--max-sentences/--batch-size cannot be larger than --buffer-size'

    print(args)

    use_cuda = torch.cuda.is_available() and not args.cpu

    # Setup task, e.g., translation

def get_parser(desc, default_task='translation'):
    # Before creating the true parser, we need to import optional user module
    # in order to eagerly import custom tasks, optimizers, architectures, etc.
    usr_parser = argparse.ArgumentParser(add_help=False, allow_abbrev=False)
    usr_parser.add_argument('--user-dir', default=None)
    usr_args, _ = usr_parser.parse_known_args()
    utils.import_user_module(usr_args)

    parser = argparse.ArgumentParser(allow_abbrev=False)
    # fmt: off
    parser.add_argument('--no-progress-bar', action='store_true', help='disable progress bar')
    parser.add_argument('--log-interval', type=int, default=1000, metavar='N',
                        help='log progress every N batches (when progress bar is disabled)')
    parser.add_argument('--log-format', default=None, help='log format to use',
                        choices=['json', 'none', 'simple', 'tqdm'])
    parser.add_argument('--tensorboard-logdir', metavar='DIR', default='',
                        help='path to save logs for tensorboard, should match --logdir '
                             'of running tensorboard (default: no tensorboard logging)')
    parser.add_argument('--seed', default=1, type=int, metavar='N',
                        help='pseudo random number generator seed')
    parser.add_argument('--cpu', action='store_true', help='use CPU instead of CUDA')
    parser.add_argument('--fp16', action='store_true', help='use FP16')
    parser.add_argument('--memory-efficient-fp16', action='store_true',

def main(args):
    utils.import_user_module(args)

    print(args)

    os.makedirs(args.destdir, exist_ok=True)
    target = not args.only_source

    task = tasks.get_task(args.task)

    def train_path(lang):
        return "{}{}".format(args.trainpref, ("." + lang) if lang else "")

    def file_name(prefix, lang):
        fname = prefix
        if lang is not None:
            fname += ".{lang}".format(lang=lang)
        return fname

def main(args):
    assert args.path is not None, '--path required for recognition!'
    assert not args.sampling or args.nbest == args.beam, \
        '--sampling requires --nbest to be equal to --beam'

    utils.import_user_module(args)

    if args.max_tokens is None and args.max_sentences is None:
        args.max_tokens = 12000
    print(args)

    use_cuda = torch.cuda.is_available() and not args.cpu

    # Load dataset split
    task = tasks.setup_task(args)
    task.load_dataset(args.gen_subset)

    # Set dictionary
    dictionary = task.target_dictionary

    # Load ensemble
    print('| loading model(s) from {}'.format(args.path))