ilsp/flores200_en-el

---
language:
- en
- el
license: cc-by-sa-4.0
size_categories:
- 1K<n<10K
task_categories:
- translation
dataset_info:
features:
- name: en
dtype: string
- name: el
dtype: string
splits:
- name: validation
num_bytes: 406555
num_examples: 997
- name: test
num_bytes: 427413
num_examples: 1012
download_size: 481524
dataset_size: 833968
configs:
- config_name: default
data_files:
- split: validation
path: data/validation-*
- split: test
path: data/test-*
---
# FLORES-200 EN-EL with prompts for translation by LLMs
Based on [FLORES-200](https://huggingface.co/datasets/Muennighoff/flores200) dataset.

Publication:
@article{nllb2022,
author = {NLLB Team, Marta R. Costa-jussà, James Cross, Onur Çelebi, Maha Elbayad, Kenneth Heafield, Kevin Heffernan, Elahe Kalbassi, Janice Lam, Daniel Licht, Jean Maillard, Anna Sun, Skyler Wang, Guillaume Wenzek, Al Youngblood, Bapi Akula, Loic Barrault, Gabriel Mejia Gonzalez, Prangthip Hansanti, John Hoffman, Semarley Jarrett, Kaushik Ram Sadagopan, Dirk Rowe, Shannon Spruit, Chau Tran, Pierre Andrews, Necip Fazil Ayan, Shruti Bhosale, Sergey Edunov, Angela Fan, Cynthia Gao, Vedanuj Goswami, Francisco Guzmán, Philipp Koehn, Alexandre Mourachko, Christophe Ropers, Safiyyah Saleem, Holger Schwenk, Jeff Wang},
title = {No Language Left Behind: Scaling Human-Centered Machine Translation},
year = {2022}
}

Number of examples : 1012

## FLORES-200 for EN to EL with 0-shot prompts
Contains 2 prompt variants:
- EN:\n\[English Sentence\]\nEL:
- English:\n\[English Sentence\]\nΕλληνικά:

## FLORES-200 for EL to EN with 0-shot prompts
Contains 2 prompt variants:
- EL:\n\[Greek Sentence\]\nEL:
- Ελληνικά:\n\[Greek Sentence\]\nEnglish:

## How to load datasets
```python
from datasets import load_dataset

input_file = 'flores200.en2el.test.0-shot.json'
dataset = load_dataset(
'json',
data_files=input_file,
field='examples',
split='train'
)
```

## How to generate translation results with different configurations
```python
from multiprocessing import cpu_count

def generate_translations(datapoint, config, config_name):
for idx, variant in enumerate(datapoint["prompts_results"]):
# REPLACE generate WITH ACTUAL FUNCTION WHICH TAKES GENERATION CONFIG
result = generate(variant["prompt"], config=config)
datapoint["prompts_results"][idx].update({config_name: result})
return datapoint

dataset = dataset.map(
function=generate_translations,
fn_kwargs={"config": config, "config_name": config_name},
keep_in_memory=False,
num_proc=min(len(dataset), cpu_count()),
)
```

## How to push updated datasets to hub
```python
from huggingface_hub import HfApi

input_file = "flores200.en2el.test.0-shot.json"
model_name = "meltemi-v0.2"
output_file = input_file.replace(".json", ".{}.json".format(model_name)

dataset.to_json(output_file,
force_ascii=False,
indent=4,
orient="index")

api = HfApi()

api.upload_file(
path_or_fileobj=output_file,
path_in_repo="results/{}/{}".format(model_name, output_file)
repo_id="ilsp/flores200-en-el-prompt",
repo_type="dataset",
)
```

原始信息汇总

数据集概述

基本信息

• 语言: 英语 (en) 和 希腊语 (el)
• 许可证: cc-by-sa-4.0
• 大小类别: 1K<n<10K
• 任务类别: 翻译

数据集结构

• 特征:
  • en: 字符串类型
  • el: 字符串类型
• 分割:
  • validation: 406555 字节, 997 个样本
  • test: 427413 字节, 1012 个样本
• 下载大小: 481524 字节
• 数据集大小: 833968 字节

配置

• 默认配置:
  • validation: 数据路径为 data/validation-*
  • test: 数据路径为 data/test-*

数据集内容

• 示例数量: 1012
• 提示变体:
  • 英语到希腊语:
    • EN: [English Sentence] EL:
    • English: [English Sentence] Ελληνικά:
  • 希腊语到英语:
    • EL: [Greek Sentence] EL:
    • Ελληνικά: [Greek Sentence] English:

加载数据集

python from datasets import load_dataset

input_file = flores200.en2el.test.0-shot.json dataset = load_dataset( json, data_files=input_file, field=examples, split=train )

生成翻译结果

python from multiprocessing import cpu_count

def generate_translations(datapoint, config, config_name): for idx, variant in enumerate(datapoint["prompts_results"]): result = generate(variant["prompt"], config=config) datapoint["prompts_results"][idx].update({config_name: result}) return datapoint

dataset = dataset.map( function=generate_translations, fn_kwargs={"config": config, "config_name": config_name}, keep_in_memory=False, num_proc=min(len(dataset), cpu_count()), )

推送更新数据集

python from huggingface_hub import HfApi

input_file = "flores200.en2el.test.0-shot.json" model_name = "meltemi-v0.2" output_file = input_file.replace(".json", ".{}.json".format(model_name)

dataset.to_json(output_file, force_ascii=False, indent=4, orient="index")

api = HfApi()

api.upload_file( path_or_fileobj=output_file, path_in_repo="results/{}/{}".format(model_name, output_file) repo_id="ilsp/flores200-en-el-prompt", repo_type="dataset", )