Update README.md
Browse files
README.md
CHANGED
@@ -34,7 +34,7 @@ configs:
|
|
34 |
- **Diverse speakers**: 2296 speakers (60% male, 40% female)
|
35 |
- **Multi-source**: Derived from [VoxCeleb](https://www.robots.ox.ac.uk/~vgg/data/voxceleb/) and [Expresso](https://speechbot.github.io/expresso/) corpora
|
36 |
- **Rich metadata**: Emotion labels, NV annotations, speaker IDs, audio quality metrics
|
37 |
-
|
38 |
<!-- ## Dataset Structure π
|
39 |
|
40 |
|
@@ -48,7 +48,7 @@ NonverbalTTS/
|
|
48 |
βββ metadata.csv # Metadata annotations -->
|
49 |
|
50 |
|
51 |
-
## Metadata Schema (`metadata.csv`) π
|
52 |
|
53 |
| Column | Description | Example |
|
54 |
|--------|-------------|---------|
|
@@ -62,16 +62,16 @@ NonverbalTTS/
|
|
62 |
| `duration` | Audio length (seconds) | `3.6338125` |
|
63 |
| `speaker_id` | Speaker identifier | `ex01` |
|
64 |
| `data_name` | Source corpus | `Expresso` |
|
65 |
-
| `gender` | Speaker gender | `m` |
|
66 |
|
67 |
-
**NV Symbols**: π¬οΈ=Breath, π=Laughter, etc. (See [Annotation Guidelines](https://zenodo.org/records/15274617))
|
68 |
|
69 |
## Loading the Dataset π»
|
70 |
|
71 |
```python
|
72 |
from datasets import load_dataset
|
73 |
|
74 |
-
dataset = load_dataset("deepvk/NonverbalTTS")
|
75 |
```
|
76 |
|
77 |
<!-- # Access train split
|
@@ -122,6 +122,8 @@ Fine-tuning CosyVoice-300M on NonverbalTTS achieves parity with state-of-the-art
|
|
122 |
|
123 |
## Citation π
|
124 |
|
|
|
|
|
125 |
```
|
126 |
@dataset{nonverbaltts2024,
|
127 |
author = {Borisov Maksim, Spirin Egor, Dyatlova Darya},
|
@@ -133,4 +135,4 @@ Fine-tuning CosyVoice-300M on NonverbalTTS achieves parity with state-of-the-art
|
|
133 |
doi = {10.5281/zenodo.15274617},
|
134 |
url = {https://zenodo.org/records/15274617}
|
135 |
}
|
136 |
-
```
|
|
|
34 |
- **Diverse speakers**: 2296 speakers (60% male, 40% female)
|
35 |
- **Multi-source**: Derived from [VoxCeleb](https://www.robots.ox.ac.uk/~vgg/data/voxceleb/) and [Expresso](https://speechbot.github.io/expresso/) corpora
|
36 |
- **Rich metadata**: Emotion labels, NV annotations, speaker IDs, audio quality metrics
|
37 |
+
- **Sampling rate**: 16kHz sampling rate for audio from VoxCeleb, 48kHz for audio from Expresso
|
38 |
<!-- ## Dataset Structure π
|
39 |
|
40 |
|
|
|
48 |
βββ metadata.csv # Metadata annotations -->
|
49 |
|
50 |
|
51 |
+
<!-- ## Metadata Schema (`metadata.csv`) π
|
52 |
|
53 |
| Column | Description | Example |
|
54 |
|--------|-------------|---------|
|
|
|
62 |
| `duration` | Audio length (seconds) | `3.6338125` |
|
63 |
| `speaker_id` | Speaker identifier | `ex01` |
|
64 |
| `data_name` | Source corpus | `Expresso` |
|
65 |
+
| `gender` | Speaker gender | `m` | -->
|
66 |
|
67 |
+
<!-- **NV Symbols**: π¬οΈ=Breath, π=Laughter, etc. (See [Annotation Guidelines](https://zenodo.org/records/15274617)) -->
|
68 |
|
69 |
## Loading the Dataset π»
|
70 |
|
71 |
```python
|
72 |
from datasets import load_dataset
|
73 |
|
74 |
+
dataset = load_dataset("deepvk/NonverbalTTS", revision="refs/convert/parquet")
|
75 |
```
|
76 |
|
77 |
<!-- # Access train split
|
|
|
122 |
|
123 |
## Citation π
|
124 |
|
125 |
+
TODO
|
126 |
+
<!--
|
127 |
```
|
128 |
@dataset{nonverbaltts2024,
|
129 |
author = {Borisov Maksim, Spirin Egor, Dyatlova Darya},
|
|
|
135 |
doi = {10.5281/zenodo.15274617},
|
136 |
url = {https://zenodo.org/records/15274617}
|
137 |
}
|
138 |
+
``` -->
|