|
--- |
|
dataset_info: |
|
features: |
|
- name: id |
|
dtype: int64 |
|
- name: text |
|
dtype: string |
|
- name: language |
|
dtype: string |
|
splits: |
|
- name: train |
|
num_bytes: 60314308571 |
|
num_examples: 18000000 |
|
download_size: 34795421825 |
|
dataset_size: 60314308571 |
|
configs: |
|
- config_name: default |
|
data_files: |
|
- split: train |
|
path: data/train-* |
|
language: |
|
- en |
|
- de |
|
- es |
|
- it |
|
- fr |
|
- ru |
|
--- |
|
# OSCAR EU 6x3M Dataset |
|
|
|
## Overview |
|
The OSCAR EU 6x3M dataset is a carefully curated subset of the larger OSCAR corpus, specifically focusing on the main European languages. This dataset includes a balanced representation of six languages: English (en), German (de), Spanish (es), Italian (it), French (fr), and Russian (ru). The "6x3M" in the name signifies that each language is represented with approximately 3 million randomly sampled documents, providing a comprehensive and diverse linguistic resource. |
|
|
|
## Dataset Description |
|
- **Languages Included**: English, German, Spanish, Italian, French, Russian |
|
- **Number of Documents**: Approximately 18 million (3 million per language) |
|
- **Data Source**: The dataset is derived from the OSCAR corpus, a large multilingual corpus created from the Common Crawl. |
|
|
|
## Use Cases |
|
This dataset is ideal for a variety of natural language processing applications, including but not limited to: |
|
- Multilingual language modeling |
|
- Cross-linguistic transfer learning |
|
- Language identification and classification |
|
- Comparative linguistic studies |
|
|
|
## Accessing the Dataset |
|
The dataset is available through the HuggingFace Datasets library. You can load the dataset using the following code snippet: |
|
```python |
|
from datasets import load_dataset |
|
|
|
dataset = load_dataset("oscar_eu_6x3M") |
