---
language:
- en
- hi
- bn
- ta
- te
license: cc-by-nc-4.0
task_categories:
- translation
tags:
- parallel-corpus
- indic
- samanantar
---

# samanantar-parallel-en-hi-bn-ta-te

Row-aligned parallel sentences for **English, Hindi, Bengali, Tamil, Telugu** built from
[`ai4bharat/samanantar`](https://huggingface.co/datasets/ai4bharat/samanantar) by inner-joining
on the English ``src`` field (``idx`` is **not** shared across language configs).

## Columns

| Column | Description |
|--------|-------------|
| `en` | English (Samanantar ``src``) |
| `hi` | Hindi translation |
| `bn` | Bengali translation |
| `ta` | Tamil translation |
| `te` | Telugu translation |

## Load

```python
from datasets import load_dataset
ds = load_dataset("nirmalendu01/samanantar-parallel-en-hi-bn-ta-te", split="train")
```

## Provenance

- Source: Samanantar v0.3.0 (CC BY-NC 4.0)
- Join: exact match on stripped ``src`` string across configs ``hi``, ``bn``, ``ta``, ``te``