File size: 1,289 Bytes
7b0884d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
language: 
- ca
- jp

tags:
- translation

library_name: opennmt
license: mit
metrics:
- bleu

inference: false
---

### Introduction

Catalan - Japanese translation model for OpenNMT. These are the same models that we have in production at https://www.softcatala.org/traductor/.
The models are quantified for low latency.

### Usage

Install the necessary dependencies:


```bash
pip3 install ctranslate2 pyonmttok
```


Simple tokenization & translation using Python:


```python
import ctranslate2
import pyonmttok
from huggingface_hub import snapshot_download
model_dir = snapshot_download(repo_id="softcatala/translate-cat-jpn", revision="main")

tokenizer=pyonmttok.Tokenizer(mode="none", sp_model_path = model_dir + "/sp_m.model")
tokenized=tokenizer.tokenize("Hola amics")

translator = ctranslate2.Translator(model_dir)
translated = translator.translate_batch([tokenized[0]])
print(tokenizer.detokenize(translated[0][0]['tokens']))
```

## Benchmarks

| testset                               | BLEU  |
|---------------------------------------|-------|
| test dataset (from train/dev/test)   	| 21.3	|
| Flores200 dataset                     | 19.8	|

## Additional information
* https://github.com/Softcatala/nmt-models
* https://github.com/Softcatala/parallel-catalan-corpus