andreiaalexa/wmt18-cs-en-preprocessed

---
language:
- cs
- en
task_categories:
- translation
pretty_name: WMT18 Czech-English Preprocessed
size_categories:
- 10K<n<100K
---

# WMT18 Czech-English Preprocessed

This dataset is a preprocessed subset of the WMT18 Czech-English translation dataset.

Original dataset: https://huggingface.co/datasets/wmt/wmt18

## Dataset Description

The dataset contains Czech-English parallel sentence pairs for machine translation.  
Each example contains one Czech sentence and its corresponding English translation.

## Preprocessing

The following preprocessing steps were applied:

1. Loaded the `cs-en` configuration from `wmt/wmt18`.
2. Removed sentence pairs where either the Czech or English sentence contains fewer than 2 words.
3. Shuffled the dataset using a fixed random seed: `42`.
4. Selected fixed-size splits for training, validation, and testing.

Using a fixed seed makes the dataset split reproducible.

## Dataset Splits

| Split | Number of examples |
|---|---:|
| train | 48,000 |
| validation | 1,000 |
| test | 1,000 |

## Data Format

Each row is stored in JSONL format:

```json
{
  "translation": {
    "cs": "Czech sentence",
    "en": "English sentence"
  }
}

Usage

from datasets import load_dataset

ds = load_dataset("charlie0831/wmt18-cs-en-preprocessed")

print(ds)
print(ds["train"][0])

If the splits are not detected automatically:

from datasets import load_dataset

data_files = {
    "train": "train.jsonl",
    "validation": "validation.jsonl",
    "test": "test.jsonl",
}

ds = load_dataset(
    "charlie0831/wmt18-cs-en-preprocessed",
    data_files=data_files
)

Citation

Please cite the original WMT18 dataset if this dataset is used.

Original dataset: https://huggingface.co/datasets/wmt/wmt18