Design upgrade: Sleek tier-1 dataset card matching ViuAI-500M standards

README.md

@@ -1,8 +1,8 @@
 ---
+license: apache-2.0
 language:
 - en
 - hi
-license: apache-2.0
 task_categories:
 - translation
 size_categories:
@@ -12,6 +12,7 @@ tags:
 - translation
 - nmt
 - viuai
+- sarus-500m
 - english-hindi
 - indic
 - devanagari
@@ -30,49 +31,48 @@ configs:
 <div align="center">
 
 # 📚 ViuTranslate-Data
-### *Industry-Standard Bilingual Parallel Corpus (English ↔ Hindi)*
 
-[![Dataset Provider](https://img.shields.io/badge/Provider-ViuAI%20Research-blue.svg?style=for-the-badge&logo=huggingface)](https://huggingface.co/ViuAI)
-[![Total Pairs](https://img.shields.io/badge/Curated%20Pairs-102%2C502-green.svg?style=for-the-badge)](https://huggingface.co/datasets/ViuAI/ViuTranslate-Data)
-[![Active Tokens](https://img.shields.io/badge/Tokens-13.6%20Million-orange.svg?style=for-the-badge)](https://huggingface.co/datasets/ViuAI/ViuTranslate-Data)
-[![License](https://img.shields.io/badge/License-Apache%202.0-red.svg?style=for-the-badge)](https://www.apache.org/licenses/LICENSE-2.0)
-[![Synthetic Data](https://img.shields.io/badge/Synthetic%20Data-0%25%20(100%25%20Human)-purple.svg?style=for-the-badge)](#)
+### *A 100% Authentic Human-Curated Parallel Corpus for English ↔ Hindi Neural Translation*
 
-<p align="center">
-  <b>100% Human-Curated</b> • <b>Academic Ground-Truth</b> • <b>Pre-tokenized Memory-Mapped Shards</b> • <b>Strictly Filtered</b>
-</p>
+[![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)
 
-</div>
+[**🌐 ViuAI Studio**](https://github.com/ViuAI) | [**🤖 Model Hub**](https://huggingface.co/ViuAI/ViuTranslate) | [**📊 Data Provenance**](#-corpus-provenance--split) | [**💡 Python Usage**](#-loading-the-dataset)
 
 ---
 
-## 📖 Dataset Summary
+</div>
 
-**ViuTranslate-Data** is a production-grade English $\leftrightarrow$ Hindi parallel corpus curated specifically for fine-tuning neural machine translation foundation models like **ViuAI Sarus-500M**.
+## 📌 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 Statistics & Sourcing
+## 📊 Corpus Provenance & Split
 
-| Source Corpus | Institution | Curated Clean Pairs | Domain Diversity |
-|---|---|---|---|
-| **IIT Bombay English-Hindi Corpus** | CFILT, IIT Bombay | **50,000 Pairs** | Government, literature, science, judicial, and news texts |
-| **AI4Bharat Samanantar** | IIT Madras / AI4Bharat | **50,000 Pairs** | Large-scale verified Indian language web & publication texts |
-| **IIT Bombay Gold Benchmark Test Set**| CFILT, IIT Bombay | **2,502 Pairs** | Internationally recognized gold benchmark validation set |
-| **Total** | — | **102,502 Pairs** | **13.6 Million Active Training Tokens** |
+| 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 Filters
+## 🛡️ Rigorous Quality Control Guardrails
 
-Real-world web corpora contain significant noise, truncation, and mistranslations. To ensure only golden data enters training, the following mathematical filters were applied:
+To prevent dataset contamination, hallucinations, and vocabulary pollution, all samples were audited through strict mathematical heuristics:
 
-1. **Length Ratio Guardrail:**
+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 Identification & Character Purity:**
+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:**
@@ -82,49 +82,50 @@ Real-world web corpora contain significant noise, truncation, and mistranslation
 
 ---
 
-## 🗂️ Repository Layout
+## 📁 Repository Structure
 
 ```
 ViuTranslate-Data/
-├── README.md                          # Comprehensive Dataset Documentation
-├── metadata.json                      # Token counts & domain definitions
+├── README.md                          # Official Dataset Card
+├── metadata.json                      # Dataset Configuration & Token Counts
 │
 ├── raw/                               # Raw Curated Sentence Pairs (JSON)
-│   ├── viu_translate_100k_train.json  # 100,000 Clean Real Training Pairs (29.6 MB)
-│   └── viu_translate_val.json         # 2,502 Gold Benchmark Validation Pairs (1.18 MB)
+│   ├── 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)        # Fast Memory-Mapped Binary Arrays
+└── Pre-tokenized Shards (.npy)        # Ready-to-Train Memory-Mapped Arrays
     ├── train_tokens.npy               # Token IDs (int32, 54.5 MB)
-    ├── train_labels.npy               # Target completion labels with -100 masking (int32)
-    ├── train_offsets.npy              # Sample boundary pointers (int64, 1.6 MB)
-    ├── train_domains.npy              # Direction IDs (int32, 0.8 MB)
-    ├── val_tokens.npy                 # Validation Token IDs (int32, 2.2 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 Domain IDs (int32, 0.02 MB)
+    └── val_domains.npy                # Validation Domains (int32, 0.02 MB)
 ```
 
 ---
 
-## 💡 How to Load in Python
+## 💡 Loading the Dataset
+
+### 1. Using Hugging Face `datasets`
 
-### 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])
-# {'en': 'The sun rises in the east and sets in the west.', 'hi': 'सूरज पूर्व में उगता है और पश्चिम में डूबता है।', 'source': 'iitb_train'}
 ```
 
-### Direct Download of Binary Shards
+### 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 successfully.")
+print(f"Loaded {len(tokens):,} memory-mapped tokens.")
 ```
 
 ---
@@ -134,7 +135,7 @@ print(f"Loaded {len(tokens):,} memory-mapped tokens successfully.")
 ```bibtex
 @misc{viutranslate_data2026,
   author = {ViuAI Research Team},
-  title = {ViuTranslate-Data: A Curated 100K+ Real Academic Parallel Corpus for English-Hindi Translation},
+  title = {ViuTranslate-Data: A 100% Authentic Human-Curated Parallel Corpus for English-Hindi Translation},
   year = {2026},
   publisher = {Hugging Face},
   journal = {Hugging Face Dataset Hub},