.ipynb_checkpoints/README-checkpoint.md

---
language:
- en
- is
tags:
- translation
license: cc-by-4.0
datasets:
- quickmt/quickmt-train.is-en
model-index:
- name: quickmt-en-is
  results:
  - task:
      name: Translation eng-isl
      type: translation
      args: isl-eng
    dataset:
      name: flores101-devtest
      type: flores_101
      args: eng_Latn isl_Latn
    metrics:
    - name: BLEU
      type: bleu
      value: 25.85
    - name: CHRF
      type: chrf
      value: 53.31
    - name: COMET
      type: comet
      value: 80.89
---


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

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


## Try it on our Huggingface Space

Give it a try before downloading here: https://huggingface.co/spaces/quickmt/QuickMT-Demo


## Model Information

* Trained using [`eole`](https://github.com/eole-nlp/eole) 
* 200M parameter transformer 'big' with 8 encoder layers and 2 decoder layers
* 32k separate Sentencepiece vocabs
* Exported for fast inference to [CTranslate2](https://github.com/OpenNMT/CTranslate2) format

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-en-is ./quickmt-en-is
```

Finally use the model in python:

```python
from quickmt import Translator

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

# Translate - set beam size to 1 for faster speed (but lower quality)
sample_text = 'Dr. Ehud Ur, professor of medicine at Dalhousie University in Halifax, Nova Scotia and chair of the clinical and scientific division of the Canadian Diabetes Association cautioned that the research is still in its early days.'

t(sample_text, beam_size=5)
```

> 'Ehud Ur, prófessor í læknisfræði við Dalhousie háskólann í Halifax, Nova Scotia og formaður klínískrar og vísindalegrar deildar kanadísku sykursýkisamtakanna varaði við því að rannsóknin væri enn á fyrstu dögum.'

```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)
```

> 'Ehud Ur, prófessor í lyflækninga við Dalhousie Háskólann í Halifax, Nova Scotia, og formaður klínískrar og vísindalegrar deildar kanadísku sykursýkisamtakanna varaði við því að rannsóknarinnar væri enn á frumdögum.'

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`. A model in safetensors format to be used with `eole` is also provided.


## 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) ("eng_Latn"->"isl_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 Nvidia RTX 4070s GPU with batch size 32.

|                                  |   bleu |   chrf2 |   comet22 |   Time (s) |
|:---------------------------------|-------:|--------:|----------:|-----------:|
| quickmt/quickmt-en-is            |  25.85 |   53.31 |     80.89 |       1.75 |
| Helsinki-NLP/opus-mt-en-is       |  18.57 |   46.63 |     76.47 |       3.93 |
| facebook/nllb-200-distilled-600M |  19.8  |   47.44 |     81.45 |      26.36 |
| facebook/nllb-200-distilled-1.3B |  23.04 |   50.93 |     84.15 |      46.58 |
| facebook/m2m100_418M             |  11.93 |   37.46 |     63.74 |      22.54 |
| facebook/m2m100_1.2B             |  17.52 |   45.25 |     77.38 |      45.06 |