This tool is the open-source code for the paper "SDA: Simple Discrete Augmentation for Contrastive Sentence Representation Learning".
Before using it, you need to install SpaCy, as well as zh_core_web_md and en_core_web_sm models. Otherwise, the program will not function properly.
All the methods described in the paper are in the augmentation_tools folder.
In this context, "NEG" represents negative enhancement, while prefixes with "cn" denote Chinese enhancement. Other naming conventions are described in the paper.
run_augmentation.py is an example script for launching the augmentation process.