File size: 5,407 Bytes
33af9d7
b95afe5
dbb77b9
 
 
dda1e5e
 
 
 
 
dbb77b9
 
 
731401d
2fca166
 
 
 
 
 
33af9d7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
b95afe5
33af9d7
dda1e5e
33af9d7
947975d
dda1e5e
 
 
 
 
 
 
 
fa9ca7c
dda1e5e
33af9d7
dda1e5e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
33af9d7
 
dda1e5e
 
 
 
 
 
 
 
 
 
 
 
dbb77b9
33af9d7
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
dda1e5e
33af9d7
947975d
dda1e5e
 
 
947975d
 
dda1e5e
 
 
b86e966
 
 
 
 
 
 
 
 
 
 
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
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124

---
language: 
  - pt
thumbnail: "Portugues SBERT for the Legal Domain"
pipeline_tag: sentence-similarity
tags:
- sentence-transformers
- sentence-similarity
- transformers
datasets:
- assin
- assin2
widget:
- source_sentence: "O advogado apresentou as provas ao juíz."
  sentences:
    - "O juíz leu as provas."
    - "O juíz leu o recurso."
    - "O juíz atirou uma pedra."
  example_title: "Example 1"
model-index:
- name: BERTimbau
  results:
  - task:
      name: STS
      type: STS
    metrics:
      - name: Pearson Correlation - assin Dataset
        type: Pearson Correlation
        value: 0.76629
      - name: Pearson Correlation - assin2 Dataset
        type: Pearson Correlation
        value: 0.82357 
      - name: Pearson Correlation - stsb_multi_mt pt Dataset
        type: Pearson Correlation
        value: 0.79120
---
# rufimelo/Legal-BERTimbau-sts-large
This is a [sentence-transformers](https://www.SBERT.net) model: It maps sentences & paragraphs to a 1024 dimensional dense vector space and can be used for tasks like clustering or semantic search.
rufimelo/Legal-BERTimbau-sts-large is based on Legal-BERTimbau-large which derives from [BERTimbau](https://huggingface.co/neuralmind/bert-large-portuguese-cased) large.
It is adapted to the Portuguese legal domain and trained for STS on portuguese datasets.
## Usage (Sentence-Transformers)
Using this model becomes easy when you have [sentence-transformers](https://www.SBERT.net) installed:
```
pip install -U sentence-transformers
```
Then you can use the model like this:
```python
from sentence_transformers import SentenceTransformer
sentences = ["Isto é um exemplo", "Isto é um outro exemplo"]

model = SentenceTransformer('rufimelo/Legal-BERTimbau-sts-large')
embeddings = model.encode(sentences)
print(embeddings)
```
## Usage (HuggingFace Transformers)
```python
from transformers import AutoTokenizer, AutoModel
import torch


#Mean Pooling - Take attention mask into account for correct averaging
def mean_pooling(model_output, attention_mask):
    token_embeddings = model_output[0] #First element of model_output contains all token embeddings
    input_mask_expanded = attention_mask.unsqueeze(-1).expand(token_embeddings.size()).float()
    return torch.sum(token_embeddings * input_mask_expanded, 1) / torch.clamp(input_mask_expanded.sum(1), min=1e-9)

# Sentences we want sentence embeddings for
sentences = ['This is an example sentence', 'Each sentence is converted']

# Load model from HuggingFace Hub
tokenizer = AutoTokenizer.from_pretrained('rufimelo/Legal-BERTimbau-sts-large')
model = AutoModel.from_pretrained('rufimelo/Legal-BERTimbau-sts-large')

# Tokenize sentences
encoded_input = tokenizer(sentences, padding=True, truncation=True, return_tensors='pt')

# Compute token embeddings
with torch.no_grad():
    model_output = model(**encoded_input)
# Perform pooling. In this case, mean pooling.
sentence_embeddings = mean_pooling(model_output, encoded_input['attention_mask'])
print("Sentence embeddings:")
print(sentence_embeddings)
```
## Evaluation Results STS
| Model| Assin | Assin2|stsb_multi_mt pt|
| ---------------------------------------- | ---------- | ---------- |---------- |
| Legal-BERTimbau-sts-base| 0.71457| 0.73545 | 0.72383|
| Legal-BERTimbau-sts-base-ma| 0.74874 | 0.79532|0.82254 |
| Legal-BERTimbau-sts-base-ma-v2| 0.75481 | 0.80262|0.82178|
| Legal-BERTimbau-sts-large| 0.76629| 0.82357 | 0.79120|
| Legal-BERTimbau-sts-large-v2| 0.76299 | 0.81121|0.81726 |
| Legal-BERTimbau-sts-large-ma| 0.76195| 0.81622 | 0.82608|
| Legal-BERTimbau-sts-large-ma-v2| 0.7836| 0.8462| 0.8261| 
| Legal-BERTimbau-sts-large-ma-v3| 0.7749| 0.8470| 0.8364| 
| ---------------------------------------- | ---------- |---------- |---------- |
| BERTimbau base Fine-tuned for STS|0.78455 | 0.80626|0.82841|
| BERTimbau large Fine-tuned for STS|0.78193 | 0.81758|0.83784|
| ---------------------------------------- | ---------- |---------- |---------- |
| paraphrase-multilingual-mpnet-base-v2| 0.71457| 0.79831 |0.83999 |
| paraphrase-multilingual-mpnet-base-v2 Fine-tuned with assin(s)| 0.77641|0.79831   |0.84575 |
## Training
rufimelo/Legal-BERTimbau-sts-large is based on Legal-BERTimbau-large which derives from [BERTimbau](https://huggingface.co/neuralmind/bert-base-portuguese-cased) large.
It was trained for Semantic Textual Similarity, being submitted to a fine tuning stage with the [assin](https://huggingface.co/datasets/assin) and [assin2](https://huggingface.co/datasets/assin2) datasets.
## Full Model Architecture
```
SentenceTransformer(
  (0): Transformer({'max_seq_length': 128, 'do_lower_case': False}) with Transformer model: BertModel 
  (1): Pooling({'word_embedding_dimension': 768, 'pooling_mode_cls_token': False, 'pooling_mode_mean_tokens': True, 'pooling_mode_max_tokens': False, 'pooling_mode_mean_sqrt_len_tokens': False, 'pooling_mode_weightedmean_tokens': False, 'pooling_mode_lasttoken': False})
)
```
## Citing & Authors
If you use this work, please cite BERTimbau's work:
```bibtex
@inproceedings{souza2020bertimbau,
  author    = {F{\'a}bio Souza and
               Rodrigo Nogueira and
               Roberto Lotufo},
  title     = {{BERT}imbau: pretrained {BERT} models for {B}razilian {P}ortuguese},
  booktitle = {9th Brazilian Conference on Intelligent Systems, {BRACIS}, Rio Grande do Sul, Brazil, October 20-23 (to appear)},
  year      = {2020}
}
```