Hugging Face's logo

quickmt
/
quickmt-vi-en

Translation
English
Vietnamese
Eval Results
Model card Files Community
quickmt-vi-en
File size: 4,254 Bytes 
f453810
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
 
---
language:
- en
- vi
tags:
- translation
license: cc-by-4.0
datasets:
- quickmt/quickmt-train.vi-en
model-index:
- name: quickmt-vi-en
  results:
  - task:
      name: Translation vie-eng
      type: translation
      args: vie-eng
    dataset:
      name: flores101-devtest
      type: flores_101
      args: vie_Latn eng_Latn devtest
    metrics:
    - name: BLEU
      type: bleu
      value: 37.6
    - name: CHRF
      type: chrf
      value: 62.86
    - name: COMET
      type: comet
      value: 87.21
---


# `quickmt-vi-en` Neural Machine Translation Model 

`quickmt-vi-en` is a reasonably fast and reasonably accurate neural machine translation model for translation from `vi` into `en`.


## Model Information

* Trained using [`eole`](https://github.com/eole-nlp/eole)
* 185M parameter transformer 'big' with 8 encoder layers and 2 decoder layers
* 50k joint Sentencepiece vocabulary
* Expested for fast inference to [CTranslate2](https://github.com/OpenNMT/CTranslate2) format
* Training data: https://huggingface.co/datasets/quickmt/quickmt-train.vi-en/tree/main

See the `eole` model configuration in this repository for further details and the `eole-model` for the raw `eole` (pytorch) model.

## Usage with `quickmt`

You must install the Nvidia cuda toolkit first, if you want to do GPU inference.

Next, install the `quickmt` python library and download the model:

```bash
git clone https://github.com/quickmt/quickmt.git
pip install ./quickmt/

quickmt-model-download quickmt/quickmt-vi-en ./quickmt-vi-en
```

Finally use the model in python:

```python
from quickmt impest Translator

# Auto-detects GPU, set to "cpu" to force CPU inference
t = Translator("./quickmt-vi-en/", device="auto")

# Translate - set beam size to 1 for faster speed (but lower quality)
sample_text = 'Tiến sĩ Ehud Ur, giáo sư y khoa của Trường Đại học Dalhousie ở Halifax, Nova Scotia và là chủ tịch ban lâm sàng và khoa học của Hiệp hội Bệnh Tiểu đường Canada cảnh báo rằng nghiên cứu này chỉ mới là sự khởi đầu.'

t(sample_text, beam_size=5)
```

> 'Dr. Ehud Ur, a professor of medicine at Dalhousie University in Halifax, Nova Scotia and chair of the clinical and scientific committee of the Canadian Diabetes Association, warns that the study is just the beginning.'

```python
# Get alternative translations by sampling
# You can pass any cTranslate2 `translate_batch` arguments
t([sample_text], sampling_temperature=1.2, beam_size=1, sampling_topk=50, sampling_topp=0.9)
```

> 'Dr Ehud Ur, professor of medicine at Dalhousie College in Halifax, Nova Scotia and chair of the clinical and scientific panel of the Diabetes Association of Canada warns that this study was only the beginning.'

The model is in `ctranslate2` format, and the tokenizers are `sentencepiece`, so you can use `ctranslate2` directly instead of through `quickmt`. It is also possible  to get this model to work with e.g. [LibreTranslate](https://libretranslate.com/) which also uses `ctranslate2` and `sentencepiece`.

## Metrics

`bleu` and `chrf2` are calculated with [sacrebleu](https://github.com/mjpost/sacrebleu) on the [Flores200 `devtest` test set](https://huggingface.co/datasets/facebook/flores) ("vie_Latn"->"eng_Latn"). `comet22` with the [`comet`](https://github.com/Unbabel/COMET) library and the [default model](https://huggingface.co/Unbabel/wmt22-comet-da). "Time (s)" is the time in seconds to translate the flores-devtest dataset (1012 sentences) on an RTX 4070s GPU with batch size 32 (faster speed is possible using a larger batch size).

|                                  |   bleu |   chrf2 |   comet22 |   Time (s) |
|:---------------------------------|-------:|--------:|----------:|-----------:|
| quickmt/quickmt-vi-en            |  37.6  |   62.86 |     87.21 |       1.27 |
| Helsinki-NLP/opus-mt-vi-en       |  25.93 |   53.32 |     81.77 |       3.45 |
| facebook/nllb-200-distilled-600M |  34.99 |   60.35 |     86.13 |      24.71 |
| facebook/nllb-200-distilled-1.3B |  37.67 |   62.64 |     87.12 |      36.8  |
| facebook/m2m100_418M             |  27.63 |   55.39 |     82.79 |      17.61 |
| facebook/m2m100_1.2B             |  32.79 |   59.01 |     85.58 |      33.98 |