Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/starxsky/tacotron2-jp
Base on "tacotron2-jpanese" builded & change
https://github.com/starxsky/tacotron2-jp
deep-neural-networks gpu metal-performance-shaders pytorch tacotron tts
Last synced: 3 days ago
JSON representation
Base on "tacotron2-jpanese" builded & change
- Host: GitHub
- URL: https://github.com/starxsky/tacotron2-jp
- Owner: StarxSky
- License: bsd-3-clause
- Created: 2022-08-14T02:16:45.000Z (over 2 years ago)
- Default Branch: main
- Last Pushed: 2022-09-02T14:54:37.000Z (about 2 years ago)
- Last Synced: 2024-06-19T19:05:03.022Z (5 months ago)
- Topics: deep-neural-networks, gpu, metal-performance-shaders, pytorch, tacotron, tts
- Language: Jupyter Notebook
- Homepage: https://github.com/CjangCjengh/tacotron2-japanese
- Size: 1.52 MB
- Stars: 5
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# First of all, many thanks to the author: [@CjangCjengh](https://github.com/CjangCjengh) for the code and reference implementation! This repository is some corrections to the source project, to correct some errors, and will be improved and optimized from time to time.
* Reference: [CjangCjengh/tacotron2-japanese](https://github.com/CjangCjengh/tacotron2-japanese)
## How to use
1. Put raw Japanese texts in ./filelists
2. Put WAV files in ./wav
3. (Optional) Download NVIDIA's [pretrained model](https://drive.google.com/file/d/1c5ZTuT7J08wLUoVZ2KkUs_VdZuJ86ZqA/view?usp=sharing)
4. Open ./train.ipynb to install requirements and start training
5. Download NVIDIA's [WaveGlow model](https://drive.google.com/open?id=1rpK8CzAAirq9sWZhe9nlfvxMF1dRgFbF) or [WaveGlow model](https://sjtueducn-my.sharepoint.com/:u:/g/personal/cjang_cjengh_sjtu_edu_cn/EbyZnGnCJclGl5q_M3KGWTUBq4IIqSLiGznFdqHbv3WM5A?e=8c2aWE) based on Ayachi Nene
6. The pre-trained tacotron2 model is shown in [this](https://github.com/StarxSky/tacotron2-JP#models).
7. Open ./inference.ipynb to generate voice(Please remember to modify the corresponding path where the model is stored！！)
- ```Inference.ipynb```
```python
checkpoint_path = 'tacotron2_Model\\ayachi_nene_1'#"ayachi_nene_1"
model = load_model(hparams)
if torch.cuda.is_available() :
model.load_state_dict(torch.load(checkpoint_path)['state_dict'])
model.to(device).eval()
else :
model.load_state_dict(torch.load(checkpoint_path, map_location='cpu')['state_dict'])
model.to(device).eval()
```
- ```Inference.ipynb```
```python
waveglow_path = 'W_Model\\model' #waveglow model
if hparams.cuda_enabled :
waveglow = torch.load(waveglow_path)['model']
waveglow.cuda().eval()#.half()
else :
waveglow = torch.load(waveglow_path, map_location='cpu')['model']
waveglow.to(device).eval()#.half()
for k in waveglow.convinv:
k.float()
denoiser = Denoiser(waveglow)
```
## Cleaners
File ./hparams.py line 30
### 1. 'japanese_cleaners'
#### Before
何かあったらいつでも話して下さい。学院のことじゃなく、私事に関することでも何でも
#### After
nanikaacltaraitsudemohanashItekudasai.gakuiNnokotojanaku,shijinikaNsurukotodemonanidemo.
### 2. 'japanese_tokenization_cleaners'
#### Before
何かあったらいつでも話して下さい。学院のことじゃなく、私事に関することでも何でも
#### After
nani ka acl tara itsu demo hanashi te kudasai. gakuiN no koto ja naku, shiji nikaNsuru koto de mo naNdemo.
### 3. 'japanese_accent_cleaners'
#### Before
何かあったらいつでも話して下さい。学院のことじゃなく、私事に関することでも何でも
#### After
:na)nika a)cltara i)tsudemo ha(na)shIte ku(dasa)i.:ga(kuiNno ko(to)janaku,:shi)jini ka(Nsu)ru ko(to)demo na)nidemo.
### 4. 'japanese_phrase_cleaners'
#### Before
何かあったらいつでも話して下さい。学院のことじゃなく、私事に関することでも何でも
#### After
nanika acltara itsudemo hanashIte kudasai. gakuiNno kotojanaku, shijini kaNsuru kotodemo nanidemo.
## Models
Remember to change this line in ./inference.ipynb
```python
sequence = np.array(text_to_sequence(text, ['japanese_cleaners']))[None, :]
```
### Sanoba Witch
#### Ayachi Nene
| Cleaners Classes | Model |
| ----------- | ----------- |
| japanese_cleaners | [Model 1](https://sjtueducn-my.sharepoint.com/:u:/g/personal/cjang_cjengh_sjtu_edu_cn/ESltqOvyK3ZPsLMQwpv5FH0BoX8slLVsz3eUKwHHKkg9ww?e=vc5fdd) |
| japanese_tokenization_cleaners | [Model 2](https://sjtueducn-my.sharepoint.com/:u:/g/personal/cjang_cjengh_sjtu_edu_cn/ETNLDYH_ZRpMmNR0VGALhNQB5-LiJOqTaWQz8tXtbvCV-g?e=7nf2Ec) |
|japanese_accent_cleaners| [Model 3](https://sjtueducn-my.sharepoint.com/:u:/g/personal/cjang_cjengh_sjtu_edu_cn/Eb0WROtOsYBInTmQQZHf36IBSXmyVd4JiCF7OnQjOZkjGg?e=qbbsv4) |
#### Inaba Meguru
| Cleaners Classes | Model |
| ----------- | ----------- |
| japanese_tokenization_cleaners | [Model 1](https://sjtueducn-my.sharepoint.com/:u:/g/personal/cjang_cjengh_sjtu_edu_cn/Ed29Owd-E1NKstl_EFGZFVABe-F-a65jSAefeW_uEQuWxw?e=J628nT)|
| japanese_tokenization_cleaners | [Model 2](https://sjtueducn-my.sharepoint.com/:u:/g/personal/cjang_cjengh_sjtu_edu_cn/ER8C2tiu4-RPi_MtQ3TCuTkBVRvO1MgJOPAKpAUD4ZLiow?e=ktT81t) |
### Senren Banka
#### Tomotake Yoshino
| Cleaners Classes| Model |
| ----------- | ----------- |
| japanese_tokenization_cleaners| [Model 1](https://sjtueducn-my.sharepoint.com/:u:/g/personal/cjang_cjengh_sjtu_edu_cn/EdfFetSH3tpMr7nkiqAKzwEBXjuCRICcvgUortEvE4pdjw?e=UyvkyI)|
| japanese_phrase_cleaners| [Model 2](https://sjtueducn-my.sharepoint.com/:u:/g/personal/cjang_cjengh_sjtu_edu_cn/EeE4h5teC5xKms1VRnaNiW8BuqslFeR8VW7bCk7SWh2r8w?e=qADqbu)|
#### Murasame
| Cleaners Classes| Model |
| ----------- | ----------- |
| japanese_accent_cleaners| [Model 1](https://sjtueducn-my.sharepoint.com/:u:/g/personal/cjang_cjengh_sjtu_edu_cn/EVXUY5tNA4JOqsVL7of8GrEB4WFPrcZPRWX0MP_7G0RXfg?e=5wzBlw)|
### RIDDLE JOKER
#### Arihara Nanami
| Cleaners Classes| Model |
| ----------- | ----------- |
| japanese_accent_cleaners|[Model 1](https://sjtueducn-my.sharepoint.com/:u:/g/personal/cjang_cjengh_sjtu_edu_cn/EdxWxcjx5XdAncOdoTjtyK0BUvrigdcBb2LPmzL48q4smw?e=OlAU66)|