Checkpoint saving may be not appropriate.

[feiaa]

In your code:

        if best < test_pearson:
            best = test_pearson
            checkpoint = {
                'model': trainer.model.state_dict(), 
                'optim': trainer.optimizer,
                'pearson': test_pearson, 'mse': test_mse,
                'args': args, 'epoch': epoch
                }
            logger.debug('==> New optimum found, checkpointing everything now...')
            torch.save(checkpoint, '%s.pt' % os.path.join(args.save, args.expname))

The test_pearson is used instead of dev_pearson and the test_pearson should not be used to choose your best model. I got test result (Pearson: 0.8616 MSE: 0.2626) which had the highest dev_pearson score.