OPUS-MT-tiny-eng-deu
Distilled model from a Tatoeba-MT Teacher: OPUS-MT-models/en-de/opus-2020-02-26, 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-deu"
tokenizer = MarianTokenizer.from_pretrained(model_name)
model = MarianMTModel.from_pretrained(model_name)
tok = tokenizer("Hello, how are you?", 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+ | 36.3 | 63.8 | 0.8408 |
| Bouquet | 37.8 | 62.1 | 0.8684 |
Student
| testset | BLEU | chr-F | COMET |
|---|---|---|---|
| Flores+ | 33.2 | 61.5 | 0.8115 |
| Bouquet | 31.8 | 58.2 | 0.8260 |
Marian models
We also provide Marian-compatible versions of this model. To use them, compile Marian and run decoding with marian-decoder, for example:
marian-decoder \
-i input.txt \
-c final.model.npz.best-perplexity.npz.decoder.yml \
-m final.model.npz.best-perplexity.npz \
-v vocab.spm vocab.spm