Pixel-Linguist/rendered-stsb

---
annotations_creators:
- crowdsourced
language_creators:
- crowdsourced
- found
- machine-generated
language:
- de
- en
- es
- fr
- it
- nl
- pl
- pt
- ru
- zh
license:
- other
multilinguality:
- multilingual
size_categories:
- 10K<n<100K
source_datasets:
- extended|other-sts-b
task_categories:
- text-classification
task_ids:
- text-scoring
- semantic-similarity-scoring
pretty_name: STSb Multi MT
configs:
- config_name: default
data_files:
- path: test/*.parquet
split: test
- path: train/*.parquet
split: train
- path: dev/*.parquet
split: dev
- config_name: de
data_files:
- path: test/de.parquet
split: test
- path: train/de.parquet
split: train
- path: dev/de.parquet
split: dev
- config_name: fr
data_files:
- path: test/fr.parquet
split: test
- path: train/fr.parquet
split: train
- path: dev/fr.parquet
split: dev
- config_name: ru
data_files:
- path: test/ru.parquet
split: test
- path: train/ru.parquet
split: train
- path: dev/ru.parquet
split: dev
- config_name: zh
data_files:
- path: test/zh.parquet
split: test
- path: train/zh.parquet
split: train
- path: dev/zh.parquet
split: dev
- config_name: es
data_files:
- path: test/es.parquet
split: test
- path: train/es.parquet
split: train
- path: dev/es.parquet
split: dev
- config_name: it
data_files:
- path: test/it.parquet
split: test
- path: train/it.parquet
split: train
- path: dev/it.parquet
split: dev
- config_name: en
data_files:
- path: test/en.parquet
split: test
- path: train/en.parquet
split: train
- path: dev/en.parquet
split: dev
- config_name: pt
data_files:
- path: test/pt.parquet
split: test
- path: train/pt.parquet
split: train
- path: dev/pt.parquet
split: dev
- config_name: nl
data_files:
- path: test/nl.parquet
split: test
- path: train/nl.parquet
split: train
- path: dev/nl.parquet
split: dev
- config_name: pl
data_files:
- path: test/pl.parquet
split: test
- path: train/pl.parquet
split: train
- path: dev/pl.parquet
split: dev
---

### Dataset Summary

This dataset is rendered to images from STS-benchmark. 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 English train Dataset:
```python
from datasets import load_dataset
dataset = load_dataset("Pixel-Linguist/rendered-stsb", name="en", split="train")
```

Load Chinese dev Dataset:
```python
from datasets import load_dataset
dataset = load_dataset("Pixel-Linguist/rendered-stsb", name="zh", split="dev")
```

### Languages

de, en, es, fr, it, nl, pl, pt, ru, zh

### 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}
}
```