Efficient and Enhanced Text Summarization by Compressing and Data Augmentation for Transformers-Based Models