|
|
|
--- |
|
language: Swedish Italian |
|
tags: |
|
- translation Swedish Italian model |
|
datasets: |
|
- dcep europarl jrc-acquis |
|
widget: |
|
- text: "– med beaktande av rådet beslut om Syrien av den 12 april, 9 och 23 maj, 20 och 25 juni samt den 2 september 2011 och av uttalandena från unionens höga representant av den 9, 23 och 29 april, 9 maj, 6, 9 och 11 juni, 9 och 31 juli, 1, 4, 18 och 30 augusti samt den 2 september 2011 om en utvidgning av de restriktiva åtgärderna mot den syriska regimen," |
|
|
|
--- |
|
|
|
# legal_t5_small_trans_sv_it_small_finetuned model |
|
|
|
Model on translating legal text from Swedish to Italian. It was first released in |
|
[this repository](https://github.com/agemagician/LegalTrans). This model is first pretrained all the translation data over some unsupervised task. Then the model is trained on three parallel corpus from jrc-acquis, europarl and dcep. |
|
|
|
|
|
## Model description |
|
|
|
legal_t5_small_trans_sv_it_small_finetuned is initially pretrained on unsupervised task with the all of the data of the training set. The unsupervised task was "masked language modelling". legal_t5_small_trans_sv_it_small_finetuned is based on the `t5-small` model and was trained on a large corpus of parallel text. This is a smaller model, which scales the baseline model of t5 down by using `dmodel = 512`, `dff = 2,048`, 8-headed attention, and only 6 layers each in the encoder and decoder. This variant has about 60 million parameters. |
|
|
|
## Intended uses & limitations |
|
|
|
The model could be used for translation of legal texts from Swedish to Italian. |
|
|
|
### How to use |
|
|
|
Here is how to use this model to translate legal text from Swedish to Italian in PyTorch: |
|
|
|
```python |
|
from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline |
|
|
|
pipeline = TranslationPipeline( |
|
model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_trans_sv_it_small_finetuned"), |
|
tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_trans_sv_it", do_lower_case=False, |
|
skip_special_tokens=True), |
|
device=0 |
|
) |
|
|
|
sv_text = "– med beaktande av rådet beslut om Syrien av den 12 april, 9 och 23 maj, 20 och 25 juni samt den 2 september 2011 och av uttalandena från unionens höga representant av den 9, 23 och 29 april, 9 maj, 6, 9 och 11 juni, 9 och 31 juli, 1, 4, 18 och 30 augusti samt den 2 september 2011 om en utvidgning av de restriktiva åtgärderna mot den syriska regimen," |
|
|
|
pipeline([sv_text], max_length=512) |
|
``` |
|
|
|
## Training data |
|
|
|
The legal_t5_small_trans_sv_it_small_finetuned (the supervised task which involved only the corresponding langauge pair and as well as unsupervised task where all of the data of all language pairs were available) model was trained on [JRC-ACQUIS](https://wt-public.emm4u.eu/Acquis/index_2.2.html), [EUROPARL](https://www.statmt.org/europarl/), and [DCEP](https://ec.europa.eu/jrc/en/language-technologies/dcep) dataset consisting of 8 Million parallel texts. |
|
|
|
## Training procedure |
|
|
|
The model was trained on a single TPU Pod V3-8 for 250K steps in total, using sequence length 512 (batch size 4096). It has a total of approximately 220M parameters and was trained using the encoder-decoder architecture. The optimizer used is AdaFactor with inverse square root learning rate schedule for pre-training. |
|
|
|
### Preprocessing |
|
|
|
An unigram model trained with 88M lines of text from the parallel corpus (of all possible language pairs) to get the vocabulary (with byte pair encoding), which is used with this model. |
|
|
|
### Pretraining |
|
|
|
The pre-training data was the combined data from all the 42 language pairs. The task for the model was to predict the portions of a sentence which were masked randomly. |
|
|
|
|
|
## Evaluation results |
|
|
|
When the model is used for translation test dataset, achieves the following results: |
|
|
|
Test results : |
|
|
|
| Model | BLEU score | |
|
|:-----:|:-----:| |
|
| legal_t5_small_trans_sv_it_small_finetuned | 42.575| |
|
|
|
|
|
### BibTeX entry and citation info |
|
|
|
> Created by [Ahmed Elnaggar/@Elnaggar_AI](https://twitter.com/Elnaggar_AI) | [LinkedIn](https://www.linkedin.com/in/prof-ahmed-elnaggar/) |
|
|
