Use pre-trained models to classify between the negative and positive samples within the provided dataset object from IBM.