Datasets:
Update README.md
Browse files
README.md
CHANGED
@@ -1,3 +1,76 @@
|
|
1 |
-
---
|
2 |
-
|
3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
dataset_info:
|
3 |
+
features:
|
4 |
+
- name: id
|
5 |
+
dtype: string
|
6 |
+
- name: caption
|
7 |
+
dtype: string
|
8 |
+
- name: audio
|
9 |
+
dtype:
|
10 |
+
audio:
|
11 |
+
sampling_rate: 16000
|
12 |
+
- name: duration
|
13 |
+
dtype: float64
|
14 |
+
- name: WavPath
|
15 |
+
dtype: string
|
16 |
+
splits:
|
17 |
+
- name: test
|
18 |
+
num_bytes: 3929100778.548
|
19 |
+
num_examples: 11676
|
20 |
+
download_size: 3922084600
|
21 |
+
dataset_size: 3929100778.548
|
22 |
+
configs:
|
23 |
+
- config_name: default
|
24 |
+
data_files:
|
25 |
+
- split: test
|
26 |
+
path: data/test-*
|
27 |
+
|
28 |
+
license: cc-by-nc-4.0
|
29 |
+
task_categories:
|
30 |
+
- text-to-speech
|
31 |
+
language:
|
32 |
+
- zh
|
33 |
+
tags:
|
34 |
+
- AI
|
35 |
+
- Debating
|
36 |
+
- Expressive
|
37 |
+
pretty_name: Debatts-Data
|
38 |
+
size_categories:
|
39 |
+
- 10B<n<100B
|
40 |
+
---
|
41 |
+
# Debatts-Data: The First Madarin Rebuttal Speech Dataset for Expressive Text-to-Speech Synthesis
|
42 |
+
|
43 |
+
The Debatts-Data dataset is the first Madarin rebuttal speech dataset for expressive text-to-speech synthesis. It is constructed from a vast collection of professional Madarin speech data sourced from diverse video platforms and podcasts on the Internet. The in-the-wild collection approach ensures the real and natural rebuttal speech. In addition, the dataset contains annotations of transcription, duration and style embed.
|
44 |
+
The table and chart below provide the statistic information for the dataset. For some dataset samples and more information regarding Debatts system, please visit the [Debatts project page](https://amphionspace.github.io/debatts/).
|
45 |
+
|
46 |
+
## Dataset Specifications
|
47 |
+
|
48 |
+
| Attribute | Value |
|
49 |
+
|----------------------|---------------|
|
50 |
+
| Language | ZH |
|
51 |
+
| Number of Speakers | 2,350 (est.) |
|
52 |
+
| Duration (hrs) | 111 |
|
53 |
+
| Type | Text + Speech |
|
54 |
+
| Sample Rate (kHz) | 16 |
|
55 |
+
| Recorded Method | Wild |
|
56 |
+
|
57 |
+
|
58 |
+
The JSON files in the dataset contain the following keys:
|
59 |
+
|
60 |
+
| Key | Description |
|
61 |
+
|-------------------|----------------------------------------------------------|
|
62 |
+
| `key` | Unique identifier for each sample in the dataset |
|
63 |
+
| `text` | Text transcription of the audio |
|
64 |
+
| `duration` | Duration of the audio clip in seconds |
|
65 |
+
| `language` | Language of the audio content |
|
66 |
+
| `wav_path` | Path to the corresponding WAV file |
|
67 |
+
| `prompt0_wav_path`| Path to the WAV file used as a prompt |
|
68 |
+
| `style_feature` | Style features associated with the audio sample |
|
69 |
+
## README 🔥🔥🔥
|
70 |
+
|
71 |
+
## Dataset Usage
|
72 |
+
|
73 |
+
To utilize the Debatts-Data dataset, you can download the raw audio files from the files and versions. The Debatts-Data.tar.gz contains the training data, while the Debatts-Data_test.tar.gz contains the testing data with extra speaker prompt speech.
|
74 |
+
|
75 |
+
|
76 |
+
*Please note that Debatts-Data does not own the copyright to the audio files; the copyright remains with the original owners of the videos or audio. Users are permitted to use this dataset only for non-commercial purposes under the CC BY-NC-4.0 license.*
|