metadata
dataset_info:
features:
- name: id
dtype: int64
- name: image
dtype: image
- name: label
dtype: string
splits:
- name: train
num_bytes: 456318868.038
num_examples: 5209
download_size: 488886429
dataset_size: 456318868.038
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
language:
- ru
Dataset Card for Dataset Name
This is a dataset with memes and their descriptions in russian.
Dataset Description
This dataset is made from TG channel: https://t.me/text_meme
For each meme image there is a detailed description.
Uses
Idk, I wanted to finetune CLIP for finding memes in russian, but I am not skilled enough right now. Maybe I'll come back to it later.