This project provides paraphrased texts and simple code for Hateful Memes Challenge dataset.
data_test.jsonl
consists of ALL metadata for memes in the dataset.
id
img
label
text
so do all other files starting with data_test_paraphrased_nlpaug
.
This simple project uses simple functions from nlpaug, pandas, tqdm. Therefore, one merely go to these links and install as their instructions. It works with Python 3.7
Just read the file, it's simple.