MR973/google-smol-en-ba

---
license: cc-by-4.0
task_categories:
- translation
language:
- en
- ba
tags:
- bashkir
- google
- text
- csv
- dask
- croissant
size_categories:
- 1K<n<10K
pretty_name: Google Smol to Bashkir
configs:
- config_name: smoldoc
data_files: smoldoc.csv
- config_name: smolsent
data_files: smolsent.csv
---

# Google Smol English-Bashkir Parallel Corpus (google-smol-en-ba)

## Dataset Description
**google-smol-en-ba** is an English-to-Bashkir parallel dataset created by translating the "Smol" sentences dataset originally developed by Google. This dataset aims to support Natural Language Processing (NLP), Machine Translation (MT), and Large Language Model (LLM) training for the Bashkir language (башҡорт теле).

The translation was performed directly from the original English source text into Bashkir, and was manually curated and verified by a native speaker.

- **Source Language:** English (`en`)
- **Target Language:** Bashkir (`ba`)
- **Original Dataset:** Google Smol
- **License:** CC-BY-4.0

## Датасет тураһында
Был датасет Google корпорацияһының "Smol" проекты нигеҙендә төҙөлгән инглизсә-башҡортса параллель текст корпусы. Текстар инглиз теленән туранан-тура башҡорт теленә тәржемә ителде һәм тулыһынса кеше тарафынан тикшерелеп, төҙәтелеп сыҡты. Был мәғлүмәт базаһы башҡорт теле өсөн машина тәржемәһен һәм нейроселтәрҙе үҫтереү маҡсатында булдырылды.

## Acknowledgments
This dataset is a derived work based on the Google Smol project.
- **Original Smol Dataset Authors:** Isaac Caswell (Google), Ali Kuzhuget.
- **Bashkir Translation and Curation:** MR973

## Intended Uses
This dataset can be effectively used for:
- Training and fine-tuning machine translation models (English <-> Bashkir).
- Improving Bashkir language understanding in Large Language Models.
- Cross-lingual linguistic research.