|
|
|
--- |
|
language: English Cszech |
|
tags: |
|
- translation English Cszech model |
|
datasets: |
|
- dcep europarl jrc-acquis |
|
widget: |
|
- text: "1 In the countries concerned, this certainly affects the priority assigned to making progress on the issue of final disposal, particularly of highly radioactive waste and irradiated fuel elements." |
|
|
|
--- |
|
|
|
# legal_t5_small_trans_en_cs model |
|
|
|
Model on translating legal text from English to Cszech. It was first released in |
|
[this repository](https://github.com/agemagician/LegalTrans). This model is trained on three parallel corpus from jrc-acquis, europarl and dcep. |
|
|
|
|
|
## Model description |
|
|
|
legal_t5_small_trans_en_cs 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 English to Cszech. |
|
|
|
### How to use |
|
|
|
Here is how to use this model to translate legal text from English to Cszech in PyTorch: |
|
|
|
```python |
|
from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline |
|
|
|
pipeline = TranslationPipeline( |
|
model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_trans_en_cs"), |
|
tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_trans_en_cs", do_lower_case=False, |
|
skip_special_tokens=True), |
|
device=0 |
|
) |
|
|
|
en_text = "1 In the countries concerned, this certainly affects the priority assigned to making progress on the issue of final disposal, particularly of highly radioactive waste and irradiated fuel elements." |
|
|
|
pipeline([en_text], max_length=512) |
|
``` |
|
|
|
## Training data |
|
|
|
The legal_t5_small_trans_en_cs 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 5 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 |
|
|
|
|
|
|
|
## Evaluation results |
|
|
|
When the model is used for translation test dataset, achieves the following results: |
|
|
|
Test results : |
|
|
|
| Model | BLEU score | |
|
|:-----:|:-----:| |
|
| legal_t5_small_trans_en_cs | 50.177| |
|
|
|
|
|
### 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/) |
|
|
