|
--- |
|
license: mit |
|
language: |
|
- ru |
|
- vi |
|
metrics: |
|
- bleu |
|
pipeline_tag: translation |
|
--- |
|
|
|
## RuVi-Translation |
|
|
|
Vietnamese-Russian and Russian-Vietnamese Translation models. |
|
|
|
- **Developed by:** Minggz |
|
|
|
|
|
## How to Use |
|
```python |
|
import ctranslate2 |
|
import pyonmttok |
|
|
|
text = "Всем привет, я NLP-инженер" |
|
ruvi_model_dir = 'ruvi_ctranslate2' |
|
#viru_model_dir = 'viru_ctranslate2' |
|
|
|
ru_tokenizer=pyonmttok.Tokenizer(mode="none", sp_model_path = 'ru_spm.model') |
|
vi_tokenizer=pyonmttok.Tokenizer(mode="none", sp_model_path = 'vi_spm.model') |
|
|
|
tokenized=ru_tokenizer.tokenize(text) |
|
|
|
translator = ctranslate2.Translator(ruvi_model_dir) |
|
translated = translator.translate_batch([tokenized[0]]) |
|
print(vi_tokenizer.detokenize(translated[0][0]['tokens'])) |
|
|
|
|
|
|
