README.md

---
language:
- ro
- en
tags:
- translation
license: apache-2.0
datasets:
- yhavinga/ccmatrix
- Helsinki-NLP/tatoeba_mt
pipeline_tag: translation
---

# English-Romanian Translate
![v1.0](https://img.shields.io/badge/V.1-31.03.2023-brightgreen)
![v2.0](https://img.shields.io/badge/V.2-05.05.2023-brightgreen)
![v3.0](https://img.shields.io/badge/V.3-19.06.2023-brightgreen)

# Finetune
This model is a finetune of the Helsinki-NLP/opus-mt-en-ro model, on ccmatrix(full) dataset.

### How to use
```python
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM

tokenizer = AutoTokenizer.from_pretrained("BlackKakapo/opus-mt-en-ro")
model = AutoModelForSeq2SeqLM.from_pretrained("BlackKakapo/opus-mt-en-ro")
```

## Benchmarks
| Model | Test Dataset | BLEU | Translation Length | Reference Length |
| ------ | ------ | ------ | ------ | ------ |
| **BlackKakapo/opus-mt-en-ro** | ("Helsinki-NLP/tatoeba_mt", "eng-ron", "test") | **0.53118** | 40516 | 40352 |
| Helsinki-NLP/opus-mt-en-ro | ("Helsinki-NLP/tatoeba_mt", "eng-ron", "test") | 0.45844 | 39222 | 40352 |
| Helsinki-NLP/opus-tatoeba-en-ro | ("Helsinki-NLP/tatoeba_mt", "eng-ron", "test") | 0.46994 | 39423 | 40352 |