kosti4ka, updated
🕥
2022-01-21 13:47:03
Ukrainian G2P model
Installation
pip install -U ukro-g2p
Example inference
```python from ukro_g2p.predict import G2P
g2p = G2P('ukro-base-uncased')
ARPABET format
g2p('фонетика')
human readable format
g2p('фонетика', human_readable=True) ```
Jupyter notebook with the example: 
Web app
https://ukro-g2p.herokuapp.com
Code for the web app: https://github.com/kosti4ka/ukro_g2p_demo
Ukrainian phonology symbols
Голосні
| Ukrainian | ARPABET-like |
|---|---|
| [і] | IY |
| [и] | IH |
| [е] | EH |
| [у] | UH |
| [о] | AO |
| [а] | AA |
| Наближення | |
| [еи] | EIH |
| [еі] | EIY |
| [ие] | IHE |
| [оу] | AOU |
| Наголос | |
| [ ́ ] | 1 |
Приголосні
| Ukrainian | ARPABET-like |
|---|---|
| [б] | B |
| [в] | V |
| [г] | H |
| [д] | D |
| [дж] | JH |
| [дз] | DZ |
| [ж] | ZH |
| [з] | Z |
| [й] | Y |
| [к] | K |
| [л] | L |
| [м] | M |
| [н] | N |
| [п] | P |
| [р] | R |
| [с] | S |
| [т] | T |
| [ф] | F |
| [х] | X |
| [ц] | TS |
| [ч] | CH |
| [ш] | SH |
| [ґ] | G |
| [ў] | WH |
| М'які | |
| [д’] | DJ |
| [дз’] | DZJ |
| [з’] | ZJ |
| [л’] | LJ |
| [н’] | NJ |
| [р’] | RJ |
| [с’] | SJ |
| [т’] | TJ |
| [ц’] | TSJ |
| Пом'якшення | |
| [`] | 2 |
| Подовження | |
| [:] | 3 |
Issues
Цікавить питання
opened on 2021-09-04 16:58:50 by tarasfrompirЩо необхідно для того щоб покращити існуючу модель ? І взагалі чи це можливо ?
Add License
opened on 2021-08-16 14:28:36 by JRMeyerVery useful code! You should add a license so people know how they might use it in their projects:
https://choosealicense.com/
З torch==1.7.1 не працює, працює з версією torch==1.6.0
opened on 2021-02-01 22:50:55 by vldmr-dЗ версією torch==1.7.1 видає помилку RuntimeError: 'lengths' argument should be a 1D CPU int64 tensor, but got 1D cuda:0 Long tensor
Add alphabet
opened on 2020-10-06 18:06:32 by kosti4kaNeed to add alphabet to the train dataset