Update README.md
Browse files
README.md
CHANGED
|
@@ -48,9 +48,22 @@ Please notice that we encourage you to read our tutorials and learn more about
|
|
| 48 |
### Perform Text-to-Speech (TTS)
|
| 49 |
|
| 50 |
```
|
|
|
|
| 51 |
from speechbrain.pretrained import Tacotron2
|
| 52 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
mel_output, mel_length, alignment = tacotron2.encode_text("Mary had a little lamb")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 54 |
```
|
| 55 |
|
| 56 |
If you want to generate multiple sentences in one-shot, you can do in this way:
|
|
|
|
| 48 |
### Perform Text-to-Speech (TTS)
|
| 49 |
|
| 50 |
```
|
| 51 |
+
import torchaudio
|
| 52 |
from speechbrain.pretrained import Tacotron2
|
| 53 |
+
from speechbrain.pretrained import HIFIGAN
|
| 54 |
+
|
| 55 |
+
# Intialize TTS (tacotron2) and Vocoder (HiFIGAN)
|
| 56 |
+
tacotron2 = Tacotron2.from_hparams(source="speechbrain/TTS_Tacotron2", savedir="tmpdir_tts")
|
| 57 |
+
hifi_gan = HIFIGAN.from_hparams(source="speechbrain/Vocoder_HiFIGAN", savedir="tmpdir_vocoder")
|
| 58 |
+
|
| 59 |
+
# Running the TTS
|
| 60 |
mel_output, mel_length, alignment = tacotron2.encode_text("Mary had a little lamb")
|
| 61 |
+
|
| 62 |
+
# Running Vocoder (spectrogram-to-waveform)
|
| 63 |
+
waveforms = hifi_gan.decode_batch(mel_output)
|
| 64 |
+
|
| 65 |
+
# Save the waverform
|
| 66 |
+
torchaudio.save('example_TTS.wav',waveforms.squeeze(1), 22050)
|
| 67 |
```
|
| 68 |
|
| 69 |
If you want to generate multiple sentences in one-shot, you can do in this way:
|