Datasets:

Modalities:
Audio
Text
Formats:
parquet
Languages:
Arabic
ArXiv:
DOI:
Libraries:
Datasets
Dask
License:
herwoww commited on
Commit
8f31780
·
verified ·
1 Parent(s): ce66300

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +33 -3
README.md CHANGED
@@ -10,13 +10,13 @@ dataset_info:
10
  dtype: string
11
  splits:
12
  - name: train
13
- num_bytes: 277452582.0
14
  num_examples: 907
15
  - name: test
16
- num_bytes: 32977207.0
17
  num_examples: 100
18
  download_size: 302126264
19
- dataset_size: 310429789.0
20
  - config_name: Synthetic
21
  features:
22
  - name: file_name
@@ -77,4 +77,34 @@ task_categories:
77
  - automatic-speech-recognition
78
  language:
79
  - ar
 
 
 
80
  ---
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
10
  dtype: string
11
  splits:
12
  - name: train
13
+ num_bytes: 277452582
14
  num_examples: 907
15
  - name: test
16
+ num_bytes: 32977207
17
  num_examples: 100
18
  download_size: 302126264
19
+ dataset_size: 310429789
20
  - config_name: Synthetic
21
  features:
22
  - name: file_name
 
77
  - automatic-speech-recognition
78
  language:
79
  - ar
80
+ pretty_name: arvoice
81
+ size_categories:
82
+ - 10K<n<100K
83
  ---
84
+
85
+ <h2 align="center">
86
+ <b>ArVoice: A Multi-Speaker Dataset for Arabic Speech Synthesis</b>
87
+ </h2>
88
+
89
+ <div style="font-size: 16px; text-align: justify;">
90
+ <p>ArVoice is a multi-speaker Modern Standard Arabic (MSA) speech corpus with fully diacritized transcriptions, intended for multi-speaker speech synthesis, and can be useful for other tasks such as speech-based diacritic restoration, voice conversion, and deepfake detection. ArVoice comprises: (1) a new professionally recorded set from 6 voice talents with diverse demographics, (2) a modified subset of the Arabic Speech Corpus; and (3) high-quality synthetic speech from 2 commercial systems. The complete corpus consists of a total of 83.52 hours of speech across 11 voices; around 10 hours consist of human voices from 7 speakers.The modified subset and full synthetic subset are available in this repo. To access the new professionally recorded subset, <a href="/"> sign this agreement</a>. If you use the dataset or transcriptions provided in Huggingface, <u>place cite the paper</u>.
91
+ </p>
92
+ </div>
93
+
94
+ Usage Example
95
+
96
+ ```python
97
+ df = load_dataset(path="herwoww/ArVoice", data_dir="Human_3") #data_dir options: Human_3, Synthetic,
98
+ print(df)
99
+
100
+ DatasetDict({
101
+ train: Dataset({
102
+ features: ['audio', 'transcription', 'speaker_id'],
103
+ num_rows: 907
104
+ })
105
+ test: Dataset({
106
+ features: ['audio', 'transcription', 'speaker_id'],
107
+ num_rows: 100
108
+ })
109
+ })
110
+ ```