File size: 3,055 Bytes
86b09df f37e81a 86b09df f37e81a 86b09df 0358cb9 86b09df f41b673 86b09df 0358cb9 86b09df f37e81a 86b09df 1dc10a7 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 |
---
language: Swedish Cszech
tags:
- translation Swedish Cszech model
datasets:
- dcep europarl jrc-acquis
widget:
- text: "En kvalitetscertifiering av administrativa förfaranden i enlighet med ISO eller motsvarande normer skulle dessutom leda till likvärdiga villkor för sjöfartsadministrationer."
---
# legal_t5_small_trans_sv_cs model
Model on translating legal text from Swedish 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_sv_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 Swedish to Cszech.
### How to use
Here is how to use this model to translate legal text from Swedish to Cszech in PyTorch:
```python
from transformers import AutoTokenizer, AutoModelWithLMHead, TranslationPipeline
pipeline = TranslationPipeline(
model=AutoModelWithLMHead.from_pretrained("SEBIS/legal_t5_small_trans_sv_cs"),
tokenizer=AutoTokenizer.from_pretrained(pretrained_model_name_or_path = "SEBIS/legal_t5_small_trans_sv_cs", do_lower_case=False,
skip_special_tokens=True),
device=0
)
sv_text = "En kvalitetscertifiering av administrativa förfaranden i enlighet med ISO eller motsvarande normer skulle dessutom leda till likvärdiga villkor för sjöfartsadministrationer."
pipeline([sv_text], max_length=512)
```
## Training data
The legal_t5_small_trans_sv_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_sv_cs | 45.569|
### 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/)
|