File size: 1,943 Bytes
0c17a1c
299de35
1639197
 
 
 
 
 
 
 
299de35
1639197
 
0c17a1c
b8a3f38
10564d2
 
65a8b8b
10564d2
299de35
10564d2
 
 
 
6d9006b
10564d2
 
 
 
 
 
 
 
 
 
 
 
 
 
299de35
10564d2
 
 
1b67d49
10564d2
 
 
 
 
 
 
 
 
 
 
 
 
1b67d49
10564d2
1b67d49
 
 
10564d2
 
 
 
 
 
21bef79
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
---
language: "lg"
tags:
- text-to-speech
- TTS
- speech-synthesis
- Tacotron2
- speechbrain
license: "apache-2.0"
datasets:
- SALT-TTS
metrics:
- mos
---

# Sunbird AI Text-to-Speech (TTS) model trained on Luganda text 

### Text-to-Speech (TTS) with Tacotron2 trained on Professional Studio Recordings

This repository provides all the necessary tools for Text-to-Speech (TTS)  with SpeechBrain.

The pre-trained model takes in input a short text and produces a spectrogram in output. One can get the final waveform by applying a vocoder (e.g., HiFIGAN) on top of the generated spectrogram.


### Install SpeechBrain

```
pip install speechbrain
```


### Perform Text-to-Speech (TTS)

```
import torchaudio
from speechbrain.pretrained import Tacotron2
from speechbrain.pretrained import HIFIGAN

# Intialize TTS (tacotron2) and Vocoder (HiFIGAN)
tacotron2 = Tacotron2.from_hparams(source="/Sunbird/sunbird-lug-tts", savedir="tmpdir_tts")
hifi_gan = HIFIGAN.from_hparams(source="speechbrain/tts-hifigan-ljspeech", savedir="tmpdir_vocoder")

# Running the TTS
mel_output, mel_length, alignment = tacotron2.encode_text("Mbagaliza Christmass Enungi Nomwaka Omugya Gubaberere Gwamirembe")

# Running Vocoder (spectrogram-to-waveform)
waveforms = hifi_gan.decode_batch(mel_output)

# Save the waverform
torchaudio.save('example_TTS.wav',waveforms.squeeze(1), 22050)
```

If you want to generate multiple sentences in one-shot, you can do in this way:

```
from speechbrain.pretrained import Tacotron2
tacotron2 = Tacotron2.from_hparams(source="speechbrain/TTS_Tacotron2", savedir="tmpdir")

items = [
       "Nsanyuse okukulaba",
       "Erinnya lyo ggwe ani?",
       "Mbagaliza Christmass Enungi Nomwaka Omugya Gubaberere Gwamirembe"
     ]
mel_outputs, mel_lengths, alignments = tacotron2.encode_batch(items)

```

### Inference on GPU
To perform inference on the GPU, add  `run_opts={"device":"cuda"}`  when calling the `from_hparams` method.