myEarlystop.py

class EarlyStop:
    """
    Early stops the training if validation loss doesn't improve after a given patience.
    """
    def __init__(self, patience=10, verbose=False, delta=0, mode='min'):
        """
        Args:
            patience (int): How long to wait after last time validation loss improved.
                            Default: 10
            verbose (bool): If True, prints a message for each validation loss improvement.
                            Default: False
            delta (float): Minimum change in the monitored quantity to qualify as an improvement.
                            Default: 0
            mode (str): The mode of training task, `min` means the less the better, `max` means
                        the larger the better.
                        Default: min
        """
        if mode not in ['min', 'max']:
            raise ValueError('check mode')
        self.mode = mode
        self.patience = patience
        self.verbose = verbose
        self.counter = 0
        self.best_score = None
        self.early_stop = False
        self.delta = delta
        self.best_epoch = None
        if mode == 'min':
            self.init_score = float('inf')
        else:
            self.init_score = float('-inf')

    def __call__(self, score, model, save_path, epoch):
        """
        record loss or metric.
        """
        if self.mode == 'min':
            score = -score

        if self.best_score is None:
            self.best_score = score
            self.save_checkpoint(score, model, save_path)
            self.best_epoch = epoch

        elif score < self.best_score - self.delta:
            self.counter += 1
            if self.verbose:
                print('EarlyStop counter: %d out of %d' % (self.counter, self.patience))
            if self.counter >= self.patience:
                self.early_stop = True
        else:
            self.best_score = score
            self.save_checkpoint(score, model, save_path)
            self.counter = 0
            self.best_epoch = epoch

    def save_checkpoint(self, score, model, save_path):
        """
        Saves model when validation loss decrease.
        """
        if self.mode == 'min':
            if self.verbose:
                print('Validation score decreased (%.4f --> %.4f).  Saving model ...' % (self.init_score, -score))
            self.init_score = -score
        else:
            if self.verbose:
                print('Validation score increased (%.4f --> %.4f).  Saving model ...' % (self.init_score, score))
            self.init_score = score
        torch.save(model.state_dict(), save_path)