|
|
--- |
|
|
dataset_info: |
|
|
features: |
|
|
- name: en |
|
|
dtype: string |
|
|
- name: km |
|
|
dtype: string |
|
|
splits: |
|
|
- name: train |
|
|
num_bytes: 1065122 |
|
|
num_examples: 1997 |
|
|
download_size: 478172 |
|
|
dataset_size: 1065122 |
|
|
configs: |
|
|
- config_name: default |
|
|
data_files: |
|
|
- split: train |
|
|
path: data/train-* |
|
|
task_categories: |
|
|
- translation |
|
|
language: |
|
|
- km |
|
|
- en |
|
|
size_categories: |
|
|
- n<1K |
|
|
--- |
|
|
This dataset is a subset of the [`NTREX (Neural Machine Translation Test Resource for Asian Languages)`](https://github.com/MicrosoftTranslator/NTREX/tree/main), developed by Microsoft Translator. It contains parallel sentences in English (en) and Khmer (km) for machine translation research. |
|
|
|
|
|
Each entry consists of: |
|
|
- `en`: Sentence in English |
|
|
- `km`: Corresponding sentence in Khmer |
|
|
|
|
|
This resource is useful for: |
|
|
- Training or evaluating Khmer ↔ English translation models |
|
|
- Benchmarks in low-resource NMT settings |
|
|
- Fine-tuning multilingual models like mBART, NLLB, or M2M100 |
