Data Augmentation for Dialogue Summarization

SNU 2022 Fall NLP term project

Relevant files

SamSUM dataset

• imported using the datasets library provided by huggingface datasets

Files for data loading and entity-replacement (data augmentation)

• add_orig_prompts.py
• generate_entity_lists.py
• replace_entities.py
• augmented data in data folder

Files to fine-tune the T5-large language model

• pretraine model from huggingface models
• train_t5.py