umarigan's picture
Update README.md
eb95736 verified
metadata
license: creativeml-openrail-m
language:
  - tr
  - en
pretty_name: Turkish Clip Dataset
size_categories:
  - 100K<n<1M

This dataset cleaned and dowloaded version of following dataset: https://huggingface.co/datasets/visheratin/laion-coco-nllb The main purpose was to extract Turkish captions and download images. You can use this dataset to fine-tune or create a clip model. Since there English and Turkish captions you can also use those to create language model?