Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -34,7 +34,7 @@ language:
|
|
| 34 |
- twi
|
| 35 |
- yor
|
| 36 |
multilinguality:
|
| 37 |
-
- multilingual
|
| 38 |
pretty_name: Waxal NLP Datasets
|
| 39 |
task_categories:
|
| 40 |
- automatic-speech-recognition
|
|
@@ -43,49 +43,27 @@ source_datasets:
|
|
| 43 |
- UGSpeechData
|
| 44 |
- DigitalUmuganda/AfriVoice
|
| 45 |
- original
|
| 46 |
-
|
| 47 |
- config_name: asr
|
| 48 |
-
|
| 49 |
-
-
|
| 50 |
-
|
| 51 |
-
-
|
| 52 |
-
|
| 53 |
-
-
|
| 54 |
-
|
| 55 |
-
-
|
| 56 |
-
|
| 57 |
-
- name: gender
|
| 58 |
-
dtype: string
|
| 59 |
-
- name: audio
|
| 60 |
-
dtype: audio
|
| 61 |
-
splits:
|
| 62 |
-
- name: train
|
| 63 |
-
- name: validation
|
| 64 |
-
- name: test
|
| 65 |
-
- name: unlabeled
|
| 66 |
- config_name: tts
|
| 67 |
-
|
| 68 |
-
-
|
| 69 |
-
|
| 70 |
-
-
|
| 71 |
-
|
| 72 |
-
-
|
| 73 |
-
|
| 74 |
-
-
|
| 75 |
-
|
| 76 |
-
- name: gender
|
| 77 |
-
dtype: string
|
| 78 |
-
- name: creator
|
| 79 |
-
dtype: string
|
| 80 |
-
- name: file_name
|
| 81 |
-
dtype: string
|
| 82 |
-
- name: audio
|
| 83 |
-
dtype: audio
|
| 84 |
-
splits:
|
| 85 |
-
- name: train
|
| 86 |
-
- name: validation
|
| 87 |
-
- name: test
|
| 88 |
-
- name: unlabeled
|
| 89 |
---
|
| 90 |
|
| 91 |
# Waxal Datasets
|
|
@@ -124,11 +102,11 @@ consists of approximately 1,250 hours of transcribed natural speech from a wide
|
|
| 124 |
variety of voices. The 14 languages in this dataset represent over 100 million
|
| 125 |
speakers across 40 Sub-Saharan African countries.
|
| 126 |
|
| 127 |
-
|
| 128 |
-
|
| 129 |
-
|
| 130 |
-
|
| 131 |
-
|
| 132 |
|
| 133 |
### TTS Dataset
|
| 134 |
|
|
@@ -136,18 +114,19 @@ The Waxal TTS dataset is a collection of text-to-speech data in 10 African
|
|
| 136 |
languages. It consists of approximately 240 hours of scripted natural speech
|
| 137 |
from a wide variety of voices.
|
| 138 |
|
| 139 |
-
|
| 140 |
-
|
| 141 |
-
|
| 142 |
-
|
| 143 |
-
|
| 144 |
|
| 145 |
### How to Use
|
| 146 |
|
| 147 |
The `datasets` library allows you to load and pre-process your dataset in pure
|
| 148 |
Python, at scale.
|
| 149 |
|
| 150 |
-
First, ensure you have the necessary dependencies installed to handle audio
|
|
|
|
| 151 |
|
| 152 |
```bash
|
| 153 |
pip install datasets[audio]
|
|
@@ -244,10 +223,9 @@ train = tts_data['train']
|
|
| 244 |
|
| 245 |
### Data Splits
|
| 246 |
|
| 247 |
-
For the **ASR Dataset**, the data with transcriptions is split as follows:
|
| 248 |
-
*
|
| 249 |
-
*
|
| 250 |
-
* **test**: 10% of labeled data.
|
| 251 |
|
| 252 |
The **unlabeled** split contains all samples that do not have a corresponding
|
| 253 |
transcription.
|
|
@@ -259,12 +237,12 @@ The **TTS Dataset** follows a similar structure, with data split into `train`,
|
|
| 259 |
|
| 260 |
The data was gathered by multiple partners:
|
| 261 |
|
| 262 |
-
|
| 263 |
-
|
| 264 |
-
|
| 265 |
-
|
| 266 |
-
|
| 267 |
-
|
| 268 |
|
| 269 |
## Considerations for Using the Data
|
| 270 |
|
|
@@ -275,5 +253,5 @@ differ between providers.
|
|
| 275 |
|
| 276 |
## Version and Maintenance
|
| 277 |
|
| 278 |
-
-
|
| 279 |
-
-
|
|
|
|
| 34 |
- twi
|
| 35 |
- yor
|
| 36 |
multilinguality:
|
| 37 |
+
- multilingual
|
| 38 |
pretty_name: Waxal NLP Datasets
|
| 39 |
task_categories:
|
| 40 |
- automatic-speech-recognition
|
|
|
|
| 43 |
- UGSpeechData
|
| 44 |
- DigitalUmuganda/AfriVoice
|
| 45 |
- original
|
| 46 |
+
configs:
|
| 47 |
- config_name: asr
|
| 48 |
+
data_files:
|
| 49 |
+
- split: train
|
| 50 |
+
path: "data/ASR/**/train-*"
|
| 51 |
+
- split: validation
|
| 52 |
+
path: "data/ASR/**/validation-*"
|
| 53 |
+
- split: test
|
| 54 |
+
path: "data/ASR/**/test-*"
|
| 55 |
+
- split: unlabeled
|
| 56 |
+
path: "data/ASR/**/unlabeled-*"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 57 |
- config_name: tts
|
| 58 |
+
data_files:
|
| 59 |
+
- split: train
|
| 60 |
+
path: "data/TTS/**/train-*"
|
| 61 |
+
- split: validation
|
| 62 |
+
path: "data/TTS/**/validation-*"
|
| 63 |
+
- split: test
|
| 64 |
+
path: "data/TTS/**/test-*"
|
| 65 |
+
- split: unlabeled
|
| 66 |
+
path: "data/TTS/**/unlabeled-*"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 67 |
---
|
| 68 |
|
| 69 |
# Waxal Datasets
|
|
|
|
| 102 |
variety of voices. The 14 languages in this dataset represent over 100 million
|
| 103 |
speakers across 40 Sub-Saharan African countries.
|
| 104 |
|
| 105 |
+
Provider | Languages | License
|
| 106 |
+
:------------------ | :--------------------------------------- | :------------:
|
| 107 |
+
Makerere University | Acholi, Luganda, Masaaba, Nyankole, Soga | `CC-BY-4.0`
|
| 108 |
+
University of Ghana | Akan, Ewe, Dagbani, Dagaare, Ikposo | `CC-BY-NC-4.0`
|
| 109 |
+
Digital Umuganda | Fula, Lingala, Shona, Malagasy | `CC-BY-4.0`
|
| 110 |
|
| 111 |
### TTS Dataset
|
| 112 |
|
|
|
|
| 114 |
languages. It consists of approximately 240 hours of scripted natural speech
|
| 115 |
from a wide variety of voices.
|
| 116 |
|
| 117 |
+
Provider | Languages | License
|
| 118 |
+
:------------------ | :----------------------------------- | :------------:
|
| 119 |
+
Makerere University | Acholi, Luganda, Kiswahili, Nyankole | `CC-BY-4.0`
|
| 120 |
+
University of Ghana | Akan (Fante, Twi) | `CC-BY-NC-4.0`
|
| 121 |
+
Media Trust | Fula, Igbo, Hausa, Yoruba | `CC-BY-4.0`
|
| 122 |
|
| 123 |
### How to Use
|
| 124 |
|
| 125 |
The `datasets` library allows you to load and pre-process your dataset in pure
|
| 126 |
Python, at scale.
|
| 127 |
|
| 128 |
+
First, ensure you have the necessary dependencies installed to handle audio
|
| 129 |
+
data:
|
| 130 |
|
| 131 |
```bash
|
| 132 |
pip install datasets[audio]
|
|
|
|
| 223 |
|
| 224 |
### Data Splits
|
| 225 |
|
| 226 |
+
For the **ASR Dataset**, the data with transcriptions is split as follows: *
|
| 227 |
+
**train**: 80% of labeled data. * **validation**: 10% of labeled data. *
|
| 228 |
+
**test**: 10% of labeled data.
|
|
|
|
| 229 |
|
| 230 |
The **unlabeled** split contains all samples that do not have a corresponding
|
| 231 |
transcription.
|
|
|
|
| 237 |
|
| 238 |
The data was gathered by multiple partners:
|
| 239 |
|
| 240 |
+
Provider | Dataset | License
|
| 241 |
+
:------------------ | :------------------------------------------------------- | :------
|
| 242 |
+
University of Ghana | [UGSpeechData](https://doi.org/10.57760/sciencedb.22298) | `CC BY-NC-ND 4.0`
|
| 243 |
+
Digital Umuganda | [AfriVoice](DigitalUmuganda/AfriVoice) | `CC-BY 4.0`
|
| 244 |
+
Makerere University | [Yogera Dataset](https://doi.org/10.7910/DVN/BEROE0) | `CC-BY 4.0`
|
| 245 |
+
Media Trust | | `CC-BY 4.0`
|
| 246 |
|
| 247 |
## Considerations for Using the Data
|
| 248 |
|
|
|
|
| 253 |
|
| 254 |
## Version and Maintenance
|
| 255 |
|
| 256 |
+
- **Current Version:** 1.0.0
|
| 257 |
+
- **Last Updated:** 01/2026
|