Mbart Large 51 Myv Mul V1

🚀 Erzya Language Translation Model

This is a model designed to translate texts into the Erzya language (myv, cyrillic script) from 11 other languages: ru, fi, de, es, en, hi, zh, tr, uk, fr, ar. Check out its demo!

It is detailed in the paper The first neural machine translation system for the Erzya language.

✨ Features

• Multilingual Translation: Capable of translating from 11 different languages to Erzya.
• Based on mbart-large-50: Built upon facebook/mbart-large-50, with an updated vocabulary and checkpoint.
  • An extra language token myv_XX and 19K new BPE tokens are added for the Erzya language.
  • Fine - tuned for translation from Erzya: first to Russian, then to all 11 languages.

📦 Installation

There is no specific installation steps provided in the original document.

💻 Usage Examples

Basic Usage

from transformers import MBartForConditionalGeneration, MBart50Tokenizer


def fix_tokenizer(tokenizer):
    """ Add a new language token to the tokenizer vocabulary (this should be done each time after its initialization) """
    old_len = len(tokenizer) - int('myv_XX' in tokenizer.added_tokens_encoder)
    tokenizer.lang_code_to_id['myv_XX'] = old_len-1
    tokenizer.id_to_lang_code[old_len-1] = 'myv_XX'
    tokenizer.fairseq_tokens_to_ids["<mask>"] = len(tokenizer.sp_model) + len(tokenizer.lang_code_to_id) + tokenizer.fairseq_offset

    tokenizer.fairseq_tokens_to_ids.update(tokenizer.lang_code_to_id)
    tokenizer.fairseq_ids_to_tokens = {v: k for k, v in tokenizer.fairseq_tokens_to_ids.items()}
    if 'myv_XX' not in tokenizer._additional_special_tokens:
        tokenizer._additional_special_tokens.append('myv_XX')
    tokenizer.added_tokens_encoder = {}


def translate(text, model, tokenizer, src='ru_RU', trg='myv_XX', max_length='auto', num_beams=3, repetition_penalty=5.0, train_mode=False, n_out=None, **kwargs):
    tokenizer.src_lang = src
    encoded = tokenizer(text, return_tensors="pt", truncation=True, max_length=1024)
    if max_length == 'auto':
        max_length = int(32 + 1.5 * encoded.input_ids.shape[1])
    if train_mode:
        model.train()
    else:
        model.eval()
    generated_tokens = model.generate(
        **encoded.to(model.device),
        forced_bos_token_id=tokenizer.lang_code_to_id[trg], 
        max_length=max_length, 
        num_beams=num_beams,
        repetition_penalty=repetition_penalty,
        num_return_sequences=n_out or 1,
        **kwargs
    )
    out = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)
    if isinstance(text, str) and n_out is None:
        return out[0]
    return out
    

mname = 'slone/mbart-large-51-myv-mul-v1'
model = MBartForConditionalGeneration.from_pretrained(mname)
tokenizer = MBart50Tokenizer.from_pretrained(mname)
fix_tokenizer(tokenizer)


print(translate('Шумбрат, киска!', model, tokenizer, src='myv_XX', trg='ru_RU'))
# Привет, собака!   # действительно, "киска" с эрзянского переводится именно так
print(translate('Шумбрат, киска!', model, tokenizer, src='myv_XX', trg='en_XX'))
# Hi, dog!

📚 Documentation

Model Information

License

This model is licensed under cc-by-sa-4.0.

Tags

The tags associated with this model are erzya, mordovian, translation.

Supported Languages

The model supports translation from the following languages: myv, ru, fi, de, es, en, hi, zh, tr, uk, fr, ar.