README.md

metadata

dataset_info:
  features:
    - name: id
      dtype: int64
    - name: description
      dtype: string
    - name: photos
      large_list: image
  splits:
    - name: train
      num_bytes: 403295816
      num_examples: 6239
  download_size: 399709915
  dataset_size: 403295816
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: apache-2.0
task_categories:
  - image-to-text
  - visual-question-answering
language:
  - ru
multilinguality:
  - monolingual
tags:
  - meme
  - captions
  - sft
  - qa

text_meme

Соскрапено с отличного Telegram канала текстовые мемы.

Dataset Description

• Homepage https://t.me/text_meme
• Homepage https://vk.com/textmeme