sentence1
imagewidth (px)
448
448
sentence2
imagewidth (px)
448
448
score
float64
0
5
0.8
1
2.6
2.2
1.4
1.8
0.4
1.4
1
2.8
1
1.6
1.8
0.2
3.2
2
1
1.4
5
4
1
1.6
3.2
0.6
0.2
2.6
3.6
1.8
2
1
0.2
2.6
0.8
1.2
3
1.2
1
5
3.4
4.4
2.4
2.2
3.2
2.8
5
5
1.4
2.6
3.8
1.6
0.6
2
0.6
4
5
3.6
4
3.8
3.2
2
3.4
0.4
5
2.2
4.6
0.2
1.8
4
3.2
4.2
0
4.2
2
3.8
4.2
5
4.2
1.6
0
4.6
0.2
3
1.6
0.6
5
4
0.2
3.4
4.6
5
1.6
1.8
0.2
0.2
3.8
3.4
0.2
4.4
1.4
1.8

Dataset Summary

This dataset is rendered to images from STS-17. We envision the need to assess vision encoders' abilities to understand texts. A natural way will be assessing them with the STS protocols, with texts rendered into images.

Examples of Use

Load Arabic to Arabic dataset:

from datasets import load_dataset
dataset = load_dataset("Pixel-Linguist/rendered-sts17", name="ar-ar", split="test")

Load French to English dataset:

from datasets import load_dataset
dataset = load_dataset("Pixel-Linguist/rendered-sts17", name="fr-en", split="test")

Languages

ar-ar, en-ar, en-de, en-en, en-tr, es-en, es-es, fr-en, it-en, ko-ko, nl-en

Citation Information

@article{xiao2024pixel,
  title={Pixel Sentence Representation Learning},
  author={Xiao, Chenghao and Huang, Zhuoxu and Chen, Danlu and Hudson, G Thomas and Li, Yizhi and Duan, Haoran and Lin, Chenghua and Fu, Jie and Han, Jungong and Moubayed, Noura Al},
  journal={arXiv preprint arXiv:2402.08183},
  year={2024}
}
Downloads last month
60
Edit dataset card