Filtered WIT, an Image-Text Dataset.

A reliable Dataset to run Image-Text models.

You can find WIT, Wikipedia Image Text Dataset, here Data was taken from dalle-mini/wit

Author

Aarush Katta

Data Structure

The data is stored as tars, containing 10,000 samples per tar. The parquets contain the metadata of each tar, which was crated using this script Each tar contains a .jpg, .txt, and .json. The image is stored in .jpg, the caption in .txt. and the metadata in .json The preferred method to read the data is WebDataset Here's an example:

import webdataset as wds

dataset = wds.WebDataset('data/00000.tar').to_tuple('txt', 'jpg', 'json')

for text, image, meta in dataset:
    print(
      text[:50],
      image[:50],
      meta[:50]
    )

Filteration

Each sample has 8 possible captions which were compared to the image using CLIP ViT-B32 The text was encoded using multilingual CLIP text encoder Each possible caption was compared to the encoded image using Cosine Similarity and kept if the sim was greater than 0.26 Then the new caption was the filtered captions concatenated, and samples with no filtered caption were dropped. The script used is filter_wit.py