quickmt-pt-en Neural Machine Translation Model

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

Model Information

• Trained using eole
• 185M parameter transformer 'big' with 8 encoder layers and 2 decoder layers
• 50k joint Sentencepiece vocabulary
• Exported for fast inference to CTranslate2 format
• Training data: https://huggingface.co/datasets/quickmt/quickmt-train.pt-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:

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

quickmt-model-download quickmt/quickmt-pt-en ./quickmt-pt-en

Finally use the model in python:

from quickmt import Translator

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

# Translate - set beam size to 1 for faster speed (but lower quality)
sample_text = 'O Médico Ehud Ur, um professor de medicina na Universidade de Dalhousie em Halifax, Nova Escócia, e presidente da divisão clínica e científica da Canadian Diabetes Association disse que o estudo ainda está dando seus primeiros passos.'

t(sample_text, beam_size=5)

'Doctor Ehud Ur, a professor of medicine at Dalhousie University in Halifax, Nova Scotia, and president of the clinical and scientific division of the Canadian Diabetes Association said the study is still taking its first steps.'

# 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)

'Doctor Ehud Ur, a professor of medicine at Dalhousie University in Halifax, Nova Scotia, and president of the clinical and scientific division of the Canadian Diabetes Association said the study is still giving it its first steps.'

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 which also uses ctranslate2 and sentencepiece.

Metrics

bleu and chrf2 are calculated with sacrebleu on the Flores200 devtest test set ("por_Latn"->"eng_Latn"). comet22 with the comet library and the default model. "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-pt-en	48.67	71.47	89.09	1.24
Helsink-NLP/opus-mt-roa-en	45.1	69.43	88.22	3.69
facebook/nllb-200-distilled-600M	48.77	71.3	89.2	21.42
facebook/nllb-200-distilled-1.3B	51.02	72.83	89.78	36.86
facebook/m2m100_418M	40.01	66.14	86.01	17.8
facebook/m2m100_1.2B	45.69	69.52	88.3	34.42