|
--- |
|
language: |
|
- en |
|
- fr |
|
tags: |
|
- translation |
|
license: cc-by-4.0 |
|
datasets: |
|
- quickmt/quickmt-train.fr-en |
|
model-index: |
|
- name: quickmt-fr-en |
|
results: |
|
- task: |
|
name: Translation fra-eng |
|
type: translation |
|
args: fra-eng |
|
dataset: |
|
name: flores101-devtest |
|
type: flores_101 |
|
args: fra_Latn eng_Latn devtest |
|
metrics: |
|
- name: CHRF |
|
type: chrf |
|
value: 66.77 |
|
- name: BLEU |
|
type: bleu |
|
value: 42.17 |
|
- name: COMET |
|
type: comet |
|
value: 58.10 |
|
--- |
|
|
|
|
|
# `quickmt-fr-en` Neural Machine Translation Model |
|
|
|
`quickmt-fr-en` is a reasonably fast and reasonably accurate neural machine translation model for translation from `fr` 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 |
|
* Exported for fast inference to [CTranslate2](https://github.com/OpenNMT/CTranslate2) format |
|
* Training data: https://huggingface.co/datasets/quickmt/quickmt-train.fr-en/tree/main |
|
|
|
See the `eole` model configuration in this repository for further details. |
|
|
|
|
|
## 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-fr-en ./quickmt-fr-en |
|
``` |
|
|
|
Finally use the model in python: |
|
|
|
```python |
|
from quickmt import Translator |
|
|
|
# Auto-detects GPU, set to "cpu" to force CPU inference |
|
t = Translator("./quickmt-fr-en/", device="auto") |
|
|
|
# Translate - set beam size to 5 for higher quality (but slower speed) |
|
sample_text = "Résigny est une commune française située dans le département de l'Aisne, en région Hauts-de-France. " |
|
t(sample_text, beam_size=1) |
|
|
|
# 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) |
|
``` |
|
|
|
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) ("fra_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 (using `ctranslate2`) the flores-devtest dataset (1012 sentences) on an RTX 4070s GPU with batch size 32 (faster speed is possible using a large batch size). |
|
|
|
| Model | chrf2 | bleu | comet22 | Time (s) | |
|
| -------------------------------- | ----- | ------- | ------- | -------- | |
|
| quickmt/quickmt-fr-en | 68.22 | 44.28 | 88.86 | 1.1 | |
|
| Helsinki-NLP/opus-mt-fr-en | 66.85 | 41.71 | 88.31 | 3.6 | |
|
| facebook/m2m100_418M | 64.39 | 36.49 | 85.87 | 18.0 | |
|
| facebook/m2m100_1.2B | 66.51 | 41.69 | 88.00 | 34.6 | |
|
| facebook/nllb-200-distilled-600M | 67.82 | 44.04 | 88.47 | 21.7 | |
|
| facebook/nllb-200-distilled-1.3B | 69.30 | 46.22 | 89.24 | 37.1 | |
|
|
|
`quickmt-fr-en` is the fastest and is higher quality than `opus-mt-fr-en`, `m2m100_418m`, `m2m100_1.2B` and `nllb-200-distilled-600M`. |
|
|
