quickmt-en-ro Neural Machine Translation Model

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

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
• 200M parameter transformer 'big' with 8 encoder layers and 2 decoder layers
• 32k separate Sentencepiece vocabs
• Expested for fast inference to CTranslate2 format
• Training data: https://huggingface.co/datasets/quickmt/quickmt-train.lv-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-en-ro ./quickmt-en-ro

Finally use the model in python:

from quickmt import Translator

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

'Dr. Ehud Ur, professor of medicine at Dalhousie University in Halifax, Nova Scotia, and chairman of the clinical and scientific division of the Canadian Diabetic Association, pointed out that the research is still in its early stages.'

# 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, profesor de medicina la Universitatea Dalhousie din Halifax, Nova Scotia și președinte al departamentului clinic și științific al Asociației Canadiene pentru Diabet, a avertizat că cercetările au ajuns în primele zile.'

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

Metrics

bleu and chrf2 are calculated with sacrebleu on the Flores200 devtest test set ("eng_Latn"->"ron_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 Nvidia RTX 4070s GPU with batch size 32.

	bleu	chrf2	comet22	Time (s)
quickmt/quickmt-en-ro	42.29	66.07	89.67	1.48
Helsinki-NLP/opus-mt-tc-big-en-ro	40.84	65.24	87.67	3.58
facebook/nllb-200-distilled-600M	35.26	61.93	89.19	27.74
facebook/nllb-200-distilled-1.3B	37.15	63.54	90.26	48.63
facebook/m2m100_418M	32.88	59.31	84.69	22.19
facebook/m2m100_1.2B	39.35	64.16	89.25	44.61