Update README.md
Browse files
README.md
CHANGED
|
@@ -10,9 +10,25 @@ library_name: transformers
|
|
| 10 |
|
| 11 |
# NeuTTS Express
|
| 12 |
|
|
|
|
|
|
|
| 13 |
- Compact model size (0.5B)
|
| 14 |
- Emotional and paralinguistic controls via \(tags\)
|
| 15 |
- Supports streaming for low latency usecases
|
| 16 |
- Instant voice cloning
|
| 17 |
|
| 18 |
-
Supported tags: `(neutral)`, `(happy)`, `(sad)`, `(angry)`, `(surprised)`, `(disgusted)`, `(fearful)`, `(giggle)`, `(gasp)`, `(exhale)`, `(laugh)`, `(chuckle)`, `(sign)`
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
|
| 11 |
# NeuTTS Express
|
| 12 |
|
| 13 |
+
<audio controls src="https://cdn-uploads.huggingface.co/production/uploads/6599dc66eabe0f3e98de7cf6/EyqWd8BtQT-o5nyb6CavE.wav"></audio>
|
| 14 |
+
|
| 15 |
- Compact model size (0.5B)
|
| 16 |
- Emotional and paralinguistic controls via \(tags\)
|
| 17 |
- Supports streaming for low latency usecases
|
| 18 |
- Instant voice cloning
|
| 19 |
|
| 20 |
+
Supported tags: `(neutral)`, `(happy)`, `(sad)`, `(angry)`, `(surprised)`, `(disgusted)`, `(fearful)`, `(giggle)`, `(gasp)`, `(exhale)`, `(laugh)`, `(chuckle)`, `(sign)`
|
| 21 |
+
|
| 22 |
+
```python
|
| 23 |
+
from neuttsair.neutts_simple import NeuTTSAir
|
| 24 |
+
import soundfile as sf
|
| 25 |
+
|
| 26 |
+
tts = NeuTTSAir(backbone_repo="BarryFutureman/NeuTTS-Express", backbone_device="cuda",
|
| 27 |
+
codec_repo="neuphonic/distill-neucodec", codec_device="cuda")
|
| 28 |
+
input_text = "I just got the best news ever (giggle), and I couldn't be happier!"
|
| 29 |
+
start_time = time.time()
|
| 30 |
+
wav = tts.infer(input_text)
|
| 31 |
+
end_time = time.time()
|
| 32 |
+
|
| 33 |
+
sf.write("output.wav", wav, 24000)
|
| 34 |
+
```
|