Exemple #1
0
def create_trainer(args, converter):
    """
    Creates or loads a neural network according to the specified args.
    """

    logger = logging.getLogger("Logger")

    if args.load:
        logger.info("Loading provided network...")
        trainer = LmTrainer.load(args.load)
        trainer.learning_rate = args.learning_rate
    else:
        logger.info('Creating new network...')
        # sum the number of features in all extractors' tables
        input_size = converter.size() * (args.windows * 2 + 1)
        nn = LmNetwork(input_size, args.hidden, 1)
        options = {
            'learning_rate': args.learning_rate,
            'eps': args.eps,
            'ro': args.ro,
            'verbose': args.verbose,
            'left_context': args.window,
            'right_context': args.window,
            'ngram_size': args.ngrams
        }
        trainer = LmTrainer(nn, converter, options)

    trainer.saver = saver(args.output, args.vectors)

    logger.info("... with the following parameters:")
    logger.info(trainer.nn.description())

    return trainer
Exemple #2
0
def create_trainer(args, converter):
    """
    Creates or loads a neural network according to the specified args.
    """

    logger = logging.getLogger("Logger")

    if args.load:
        logger.info("Loading provided network...")
        trainer = LmTrainer.load(args.load)
        trainer.learning_rate = args.learning_rate
    else:
        logger.info('Creating new network...')
        # sum the number of features in all extractors' tables 
        input_size = converter.size() * args.windows
        nn = LmNetwork(input_size, args.hidden, 1)
        options = {
            'learning_rate': args.learning_rate,
            'eps': args.eps,
            'ro': args.ro,
            'verbose': args.verbose,
            'left_context': args.window/2,
            'right_context': args.window/2,
            'ngram_size': args.ngrams
        }
        trainer = LmTrainer(nn, converter, options)

    trainer.saver = saver(args.output, args.vectors)

    logger.info("... with the following parameters:")
    logger.info(trainer.nn.description())
    
    return trainer
Exemple #3
0
def create_trainer(args, converter):
    """
    Creates or loads a neural network according to the specified args.
    """

    logger = logging.getLogger("Logger")

    if args.load:
        logger.info("Loading provided network...")
        trainer = LmTrainer.load(args.load)
        trainer.learning_rate = args.learning_rate
    else:
        logger.info('Creating new network...')
        trainer = LmTrainer(converter,
                            args.learning_rate,
                            args.window / 2,
                            args.window / 2,
                            args.hidden,
                            ngrams=args.ngrams)

    trainer.saver = saver(args.output, args.vectors)

    logger.info("... with the following parameters:")
    logger.info(trainer.nn.description())

    return trainer
Exemple #4
0
def create_trainer(args, converter):
    """
    Creates or loads a neural network according to the specified args.
    """

    logger = logging.getLogger("Logger")

    if args.load:
        logger.info("Loading provided network...")
        trainer = LmTrainer.load(args.load)
        trainer.learning_rate = args.learning_rate
    else:
        logger.info('Creating new network...')
        trainer = LmTrainer(converter, args.learning_rate,
                            args.window/2, args.window/2,
                            args.hidden, ngrams=args.ngrams)

    trainer.saver = saver(args.output, args.vectors)

    logger.info("... with the following parameters:")
    logger.info(trainer.nn.description())
    
    return trainer