Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,139 @@
|
|
1 |
-
---
|
2 |
-
license: apache-2.0
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
license: apache-2.0
|
3 |
+
task_categories:
|
4 |
+
- text-to-audio
|
5 |
+
- text-to-speech
|
6 |
+
language:
|
7 |
+
- ur
|
8 |
+
tags:
|
9 |
+
- urdu
|
10 |
+
- tts
|
11 |
+
- g2p
|
12 |
+
pretty_name: zuhri
|
13 |
+
---
|
14 |
+
# Zuhri — Urdu G2P Dataset
|
15 |
+
|
16 |
+
**Zuhri** is a comprehensive and manually verified Urdu **Grapheme-to-Phoneme (G2P)** dataset. It is designed to aid research and development in areas such as speech synthesis, pronunciation modeling, and computational linguistics, specifically for the Urdu language.
|
17 |
+
|
18 |
+
This dataset provides accurate phoneme transcriptions and IPA representations, making it ideal for use in building high-quality TTS (Text-to-Speech), ASR (Automatic Speech Recognition), and other phonetics-related systems.
|
19 |
+
|
20 |
+
---
|
21 |
+
|
22 |
+
## 🔤 What is G2P?
|
23 |
+
|
24 |
+
Grapheme-to-Phoneme (G2P) conversion is the task of translating written text (graphemes) into spoken sounds (phonemes). G2P is crucial for generating natural-sounding speech from text and for correct pronunciation modeling in voice-driven technologies.
|
25 |
+
|
26 |
+
---
|
27 |
+
|
28 |
+
## 📁 Dataset Contents
|
29 |
+
|
30 |
+
The dataset is stored in the `data/` directory as a CSV file. Each row in the file contains the following five columns:
|
31 |
+
|
32 |
+
```
|
33 |
+
Word, Phoneme Sequence (space-separated), IPA Transcription, Verified (True/False), Frequency Rank
|
34 |
+
```
|
35 |
+
|
36 |
+
### 📋 Sample Data Preview
|
37 |
+
|
38 |
+
| Word | Phonemes | IPA | Verified | Rank |
|
39 |
+
|-------------|------------------------|--------------------|----------|------|
|
40 |
+
| عائشہ | ʕ aː ɪ ʃ ə | /ʕaːˈɪʃə/ | True | 1 |
|
41 |
+
| عطیہ | ʕ ʈ iː j ə | /ʕʈiːˈjə/ | True | 1 |
|
42 |
+
| طور | t̪ uː r | /t̪uːr/ | True | 1 |
|
43 |
+
| اردو | ʊ r d uː | /ʊrˈduː/ | True | 1 |
|
44 |
+
| اللہ | ɑ l l aː h | /ɐlˈlɑːh/ | True | 1 |
|
45 |
+
|
46 |
+
---
|
47 |
+
|
48 |
+
## 🎯 Applications
|
49 |
+
|
50 |
+
This dataset is useful for:
|
51 |
+
|
52 |
+
- Urdu Text-to-Speech (TTS) systems
|
53 |
+
- Automatic Speech Recognition (ASR)
|
54 |
+
- Phonetic alignment and transcription
|
55 |
+
- Linguistic and phonological research
|
56 |
+
- Speech technology applications in Urdu
|
57 |
+
|
58 |
+
---
|
59 |
+
|
60 |
+
## 🚀 Getting Started
|
61 |
+
|
62 |
+
### 1. Clone the Repository
|
63 |
+
|
64 |
+
```bash
|
65 |
+
git clone https://github.com/humair-m/zuhri.git
|
66 |
+
cd zuhri
|
67 |
+
```
|
68 |
+
|
69 |
+
### 2. Load the Dataset in Python
|
70 |
+
|
71 |
+
```python
|
72 |
+
import pandas as pd
|
73 |
+
|
74 |
+
# Load the Urdu G2P dataset
|
75 |
+
df = pd.read_csv("data/urdu_g2p.csv")
|
76 |
+
|
77 |
+
# Display the first few rows
|
78 |
+
print(df.head())
|
79 |
+
```
|
80 |
+
|
81 |
+
---
|
82 |
+
|
83 |
+
## 📚 Citation
|
84 |
+
|
85 |
+
If you use this dataset in your research, please cite it as follows:
|
86 |
+
|
87 |
+
### 📌 Plain Text
|
88 |
+
```
|
89 |
+
Humair Munir. URDU G2P: Zuhri Urdu Grapheme-to-Phoneme Dataset. Version 1.0, 2025. Licensed under Apache 2.0.
|
90 |
+
```
|
91 |
+
|
92 |
+
### 📌 BibTeX
|
93 |
+
```bibtex
|
94 |
+
@humair-pc{humair2025urdug2p,
|
95 |
+
author = {Humair Munir},
|
96 |
+
title = {{URDU G2P: Zuhri Urdu Grapheme-to-Phoneme Dataset}},
|
97 |
+
year = {2025},
|
98 |
+
version = {1.0},
|
99 |
+
howpublished = {\url{https://github.com/humair-m/zuhri}},
|
100 |
+
note = {Apache License 2.0}
|
101 |
+
}
|
102 |
+
```
|
103 |
+
|
104 |
+
> 📧 For academic queries, contact: **[email protected]**
|
105 |
+
|
106 |
+
---
|
107 |
+
|
108 |
+
## 📜 License
|
109 |
+
|
110 |
+
This dataset is released under the **Apache License 2.0**. You are free to use, modify, and distribute it under the terms of this license. See the `LICENSE` file for full details.
|
111 |
+
|
112 |
+
---
|
113 |
+
|
114 |
+
## 🤝 Contributing
|
115 |
+
|
116 |
+
We welcome contributions from the community! You can:
|
117 |
+
|
118 |
+
- Suggest corrections or improvements in phoneme alignments
|
119 |
+
- Add more verified Urdu words
|
120 |
+
- Extend the dataset for regional dialects
|
121 |
+
- Submit pull requests for related G2P tools or training scripts
|
122 |
+
|
123 |
+
Please open an issue to discuss your suggestions before submitting a PR.
|
124 |
+
|
125 |
+
---
|
126 |
+
|
127 |
+
## 📬 Contact
|
128 |
+
|
129 |
+
For questions, support, or collaborations, feel free to reach out:
|
130 |
+
|
131 |
+
📧 **Email**: [email protected]
|
132 |
+
|
133 |
+
---
|
134 |
+
|
135 |
+
## 🧠 Acknowledgment
|
136 |
+
|
137 |
+
This dataset is part of an ongoing effort to support low-resource languages like Urdu in the field of speech and language technologies. Thank you to the open-source community for continuous support and feedback.
|
138 |
+
|
139 |
+
---
|