ulysses115
commited on
Commit
•
da6d1dc
1
Parent(s):
fc5b11b
Upload README.md with huggingface_hub
Browse files
README.md
ADDED
@@ -0,0 +1,58 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# VITS: Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech
|
2 |
+
|
3 |
+
### Jaehyeon Kim, Jungil Kong, and Juhee Son
|
4 |
+
|
5 |
+
In our recent [paper](https://arxiv.org/abs/2106.06103), we propose VITS: Conditional Variational Autoencoder with Adversarial Learning for End-to-End Text-to-Speech.
|
6 |
+
|
7 |
+
Several recent end-to-end text-to-speech (TTS) models enabling single-stage training and parallel sampling have been proposed, but their sample quality does not match that of two-stage TTS systems. In this work, we present a parallel end-to-end TTS method that generates more natural sounding audio than current two-stage models. Our method adopts variational inference augmented with normalizing flows and an adversarial training process, which improves the expressive power of generative modeling. We also propose a stochastic duration predictor to synthesize speech with diverse rhythms from input text. With the uncertainty modeling over latent variables and the stochastic duration predictor, our method expresses the natural one-to-many relationship in which a text input can be spoken in multiple ways with different pitches and rhythms. A subjective human evaluation (mean opinion score, or MOS) on the LJ Speech, a single speaker dataset, shows that our method outperforms the best publicly available TTS systems and achieves a MOS comparable to ground truth.
|
8 |
+
|
9 |
+
Visit our [demo](https://jaywalnut310.github.io/vits-demo/index.html) for audio samples.
|
10 |
+
|
11 |
+
We also provide the [pretrained models](https://drive.google.com/drive/folders/1ksarh-cJf3F5eKJjLVWY0X1j1qsQqiS2?usp=sharing).
|
12 |
+
|
13 |
+
** Update note: Thanks to [Rishikesh (ऋषिकेश)](https://github.com/jaywalnut310/vits/issues/1), our interactive TTS demo is now available on [Colab Notebook](https://colab.research.google.com/drive/1CO61pZizDj7en71NQG_aqqKdGaA_SaBf?usp=sharing).
|
14 |
+
|
15 |
+
<table style="width:100%">
|
16 |
+
<tr>
|
17 |
+
<th>VITS at training</th>
|
18 |
+
<th>VITS at inference</th>
|
19 |
+
</tr>
|
20 |
+
<tr>
|
21 |
+
<td><img src="resources/fig_1a.png" alt="VITS at training" height="400"></td>
|
22 |
+
<td><img src="resources/fig_1b.png" alt="VITS at inference" height="400"></td>
|
23 |
+
</tr>
|
24 |
+
</table>
|
25 |
+
|
26 |
+
|
27 |
+
## Pre-requisites
|
28 |
+
0. Python >= 3.6
|
29 |
+
0. Clone this repository
|
30 |
+
0. Install python requirements. Please refer [requirements.txt](requirements.txt)
|
31 |
+
1. You may need to install espeak first: `apt-get install espeak`
|
32 |
+
0. Download datasets
|
33 |
+
1. Download and extract the LJ Speech dataset, then rename or create a link to the dataset folder: `ln -s /path/to/LJSpeech-1.1/wavs DUMMY1`
|
34 |
+
1. For mult-speaker setting, download and extract the VCTK dataset, and downsample wav files to 22050 Hz. Then rename or create a link to the dataset folder: `ln -s /path/to/VCTK-Corpus/downsampled_wavs DUMMY2`
|
35 |
+
0. Build Monotonic Alignment Search and run preprocessing if you use your own datasets.
|
36 |
+
```sh
|
37 |
+
# Cython-version Monotonoic Alignment Search
|
38 |
+
cd monotonic_align
|
39 |
+
python setup.py build_ext --inplace
|
40 |
+
|
41 |
+
# Preprocessing (g2p) for your own datasets. Preprocessed phonemes for LJ Speech and VCTK have been already provided.
|
42 |
+
# python preprocess.py --text_index 1 --filelists filelists/ljs_audio_text_train_filelist.txt filelists/ljs_audio_text_val_filelist.txt filelists/ljs_audio_text_test_filelist.txt
|
43 |
+
# python preprocess.py --text_index 2 --filelists filelists/vctk_audio_sid_text_train_filelist.txt filelists/vctk_audio_sid_text_val_filelist.txt filelists/vctk_audio_sid_text_test_filelist.txt
|
44 |
+
```
|
45 |
+
|
46 |
+
|
47 |
+
## Training Exmaple
|
48 |
+
```sh
|
49 |
+
# LJ Speech
|
50 |
+
python train.py -c configs/ljs_base.json -m ljs_base
|
51 |
+
|
52 |
+
# VCTK
|
53 |
+
python train_ms.py -c configs/vctk_base.json -m vctk_base
|
54 |
+
```
|
55 |
+
|
56 |
+
|
57 |
+
## Inference Example
|
58 |
+
See [inference.ipynb](inference.ipynb)
|