README.md

---
license: apache-2.0
language:
- en
- hi
task_categories:
- translation
size_categories:
- 100K<n<1M
tags:
- parallel-corpus
- translation
- nmt
- viuai
- sarus-500m
- english-hindi
- indic
- devanagari
- iit-bombay
- samanantar
- zero-synthetic
configs:
- config_name: default
  data_files:
  - split: train
    path: "raw/viu_translate_100k_train.json"
  - split: validation
    path: "raw/viu_translate_val.json"
---

<div align="center">

# 📚 ViuTranslate-Data

### *A 100% Authentic Human-Curated Parallel Corpus for English ↔ Hindi Neural Translation*

[![License](https://img.shields.io/badge/License-Apache%202.0-blue.svg)](https://opensource.org/licenses/Apache-2.0)
[![Dataset Pairs](https://img.shields.io/badge/Pairs-102%2C502-green.svg)](#-dataset-overview)
[![Active Tokens](https://img.shields.io/badge/Active%20Tokens-13.6%20Million-orange.svg)](#-dataset-overview)
[![Synthetic Data](https://img.shields.io/badge/Synthetic%20Data-0%25%20(Pure%20Human)-purple.svg)](#-quality-control-guardrails)
[![Languages](https://img.shields.io/badge/Languages-English%20%7C%20Hindi-red.svg)](#-dataset-overview)

[**🌐 ViuAI Studio**](https://github.com/ViuAI) | [**🤖 Model Hub**](https://huggingface.co/ViuAI/ViuTranslate) | [**📊 Data Provenance**](#-corpus-provenance--split) | [**💡 Python Usage**](#-loading-the-dataset)

---

</div>

## 📌 Introduction & Overview

**ViuTranslate-Data** is a production-grade bilingual parallel corpus curated specifically for fine-tuning neural machine translation foundation models (like **ViuAI Sarus-500M**).

Every single sentence pair is extracted from official, published academic and institutional corpora. **Zero synthetic data**, zero LLM-prompted dialogues, and zero templated variations were used.

---

## 📊 Corpus Provenance & Split

| Source Corpus | Institution | Curated Pairs | Description |
| :--- | :--- | :--- | :--- |
| **IIT Bombay English-Hindi Corpus** | CFILT, IIT Bombay | **50,000 Pairs** | Academic gold standard covering news, judicial, literature, and governmental texts |
| **AI4Bharat Samanantar** | IIT Madras / AI4Bharat | **50,000 Pairs** | Verified web & publication Indian language corpus |
| **IIT Bombay Benchmark Test Set** | CFILT, IIT Bombay | **2,502 Pairs** | Internationally accepted gold validation test suite |
| **Total Corpus** | — | **102,502 Pairs** | **13,633,518 Active Training Tokens** |

---

## 🛡️ Rigorous Quality Control Guardrails

To prevent dataset contamination, hallucinations, and vocabulary pollution, all samples were audited through strict mathematical heuristics:

1. **Length Ratio Boundary:**
   $$\quad 0.40 \le \frac{\text{len}(\text{English})}{\text{len}(\text{Hindi})} \le 2.40$$
   Pairs violating this ratio were strictly pruned to eliminate incomplete or runaway translations.
2. **Script Purity & Unicode Verification:**
   - English side enforced $\ge 50\%$ Latin alphabetic characters ($[A-Za-z]$).
   - Hindi side enforced $\ge 40\%$ Devanagari Unicode characters ($[\u0900-\u097F]$).
3. **Hygiene & Sanitation:**
   - 100% stripped of HTML tags, XML nodes, source code snippets, URLs, and file paths.
4. **Exact Cryptographic Deduplication:**
   - SHA-256 hash deduplication ensuring zero repeated pairs across train and validation sets.

---

## 📁 Repository Structure

```
ViuTranslate-Data/
├── README.md                          # Official Dataset Card
├── metadata.json                      # Dataset Configuration & Token Counts
│
├── raw/                               # Raw Curated Sentence Pairs (JSON)
│   ├── viu_translate_100k_train.json  # 100,000 Verified Academic Pairs (29.6 MB)
│   └── viu_translate_val.json         # 2,502 Gold Benchmark Pairs (1.18 MB)
│
└── Pre-tokenized Shards (.npy)        # Ready-to-Train Memory-Mapped Arrays
    ├── train_tokens.npy               # Token IDs (int32, 54.5 MB)
    ├── train_labels.npy               # Loss-Masked Target Tokens (int32, 54.5 MB)
    ├── train_offsets.npy              # Sample Boundaries (int64, 1.6 MB)
    ├── train_domains.npy              # Direction Flags (int32, 0.8 MB)
    ├── val_tokens.npy                 # Validation Tokens (int32, 2.2 MB)
    ├── val_labels.npy                 # Validation Labels (int32, 2.2 MB)
    ├── val_offsets.npy                # Validation Offsets (int64, 0.04 MB)
    └── val_domains.npy                # Validation Domains (int32, 0.02 MB)
```

---

## 💡 Loading the Dataset

### 1. Using Hugging Face `datasets`

```python
from datasets import load_dataset

dataset = load_dataset("ViuAI/ViuTranslate-Data", data_files={"train": "raw/viu_translate_100k_train.json"})
print("Sample 0:", dataset["train"][0])
```

### 2. Loading Pre-tokenized Memory-Mapped Shards (Fastest for Training)

```python
import numpy as np
from huggingface_hub import hf_hub_download

tokens_file = hf_hub_download(repo_id="ViuAI/ViuTranslate-Data", filename="train_tokens.npy", repo_type="dataset")
tokens = np.load(tokens_file, mmap_mode="r")
print(f"Loaded {len(tokens):,} memory-mapped tokens.")
```

---

## 📜 Citation

```bibtex
@misc{viutranslate_data2026,
  author = {ViuAI Research Team},
  title = {ViuTranslate-Data: A 100% Authentic Human-Curated Parallel Corpus for English-Hindi Translation},
  year = {2026},
  publisher = {Hugging Face},
  journal = {Hugging Face Dataset Hub},
  howpublished = {\url{https://huggingface.co/datasets/ViuAI/ViuTranslate-Data}}
}
```