Helsinki-NLP/opus-mt_tiny_eng-cat

OPUS-MT-tiny-eng-cat

Distilled model from a Tatoeba-MT Teacher: Tatoeba-MT-models/deu+eng+fra+por+spa-roa/opusTCv20230926max50+bt+jhubc_transformer-big_2024-05-30, which has been trained on the Tatoeba dataset.

We used the OpusDistillery to train new a new student with the tiny architecture, with a regular transformer decoder. For training data, we used Tatoeba. The configuration file fed into OpusDistillery can be found here.

How to run

from transformers import MarianMTModel, MarianTokenizer
model_name = "Helsinki-NLP/opus-mt_tiny_eng-cat"
tokenizer = MarianTokenizer.from_pretrained(model_name)
model = MarianMTModel.from_pretrained(model_name)
tok = tokenizer("Barcelona's official languages are Catalan and Spanish.", return_tensors="pt").input_ids
output = model.generate(tok)[0]
tokenizer.decode(output, skip_special_tokens=True)

Benchmarks

Teacher

testset	BLEU	chr-F	COMET
Flores+	42.6	66.8	0.8578

Student

testset	BLEU	chr-F	COMET
Flores+	40.3	65.1	0.8476