File size: 5,019 Bytes
e72dc7c
163db08
 
7df89e1
 
 
 
163db08
b3f9993
 
 
564c962
 
 
 
e72dc7c
 
b3f9993
e72dc7c
 
 
 
b18e680
163db08
 
 
d0b93db
 
 
 
 
 
83078d9
 
 
 
f091ca6
 
d0b93db
 
f091ca6
 
 
 
83078d9
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
f091ca6
b18e680
 
 
 
 
 
 
 
 
 
163db08
b18e680
163db08
 
5b91a3b
 
b18e680
 
5b91a3b
564c962
5b91a3b
564c962
 
5b91a3b
564c962
5b91a3b
 
 
564c962
 
5b91a3b
 
 
f80a6d9
5b91a3b
 
 
 
 
 
 
 
b18e680
 
 
 
 
 
5b91a3b
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
---
language: 
- en
- de
- fr
- zh
- ne
- multilingual
widget:
- text: "In December 1903 in France the Royal Swedish Academy of Sciences awarded Pierre Curie, Marie Curie, and Henri Becquerel the Nobel Prize in Physics."
- text: "Für Richard Phillips Feynman war es immer wichtig in New York, die unanschaulichen Gesetzmäßigkeiten der Quantenphysik Laien und Studenten nahezubringen und verständlich zu machen."
- text: "My name is Julian and I live in Constance"
- text: "Terence David John Pratchett est né le 28 avril 1948 à Beaconsfield dans le Buckinghamshire, en Angleterre."
- text: "北京市,通称北京(汉语拼音:Běijīng;邮政式拼音:Peking),简称“京”,是中华人民共和国的首都及直辖市,是该国的政治、文化、科技、教育、军事和国际交往中心,是一座全球城市,是世界人口第三多的城市和人口最多的首都,具有重要的国际影响力,同時也是目前世界唯一的“双奥之城”,即唯一既主办过夏季"
- text: "काठमाडौँ नेपालको सङ्घीय राजधानी र नेपालको सबैभन्दा बढी जनसङ्ख्या भएको सहर हो।"
tags:
- roberta
license: mit
datasets:
- wikiann
---

# RoBERTa for Multilingual Named Entity Recognition

## Model description

This model detects entities by classifying every token according to the IOB format:

```python
['O', 'B-PER', 'I-PER', 'B-ORG', 'I-ORG', 'B-LOC', 'I-LOC']
```

**Languages:**

TBD

## Training data

This mode was traind using a subset of all [wikiann](https://huggingface.co/datasets/wikiann) dataset.

## Evaluation results

This model achieves the following results (meassured using the validation portion of the [wikiann](https://huggingface.co/datasets/wikiann)):

| Metric      | Value |
|:-----------:|:----:|
|loss         | 87.6 |
|precision    | 87.6 |
|recall       | 87.6 |
|f1           | 87.6 |
|recall       | 87.6 |

**Per Entity Type:**

TBD

**Per Language:**

TBD


## About RoBERTa

This model is a fine-tuned version of [XLM-RoBERTa](https://huggingface.co/xlm-roberta-large). The original model was pre-trained on 2.5TB of filtered CommonCrawl data containing 100 languages. It was introduced in the paper [Unsupervised Cross-lingual Representation Learning at Scale](https://arxiv.org/abs/1911.02116) by Conneau et al. and first released in [this repository](https://github.com/pytorch/fairseq/tree/master/examples/xlmr). 

RoBERTa is a transformers model pretrained on a large corpus in a self-supervised fashion. This means it was pretrained on the raw texts only, with no humans labelling them in any way (which is why it can use lots of publicly available data) with an automatic process to generate inputs and labels from those texts.

More precisely, it was pretrained with the Masked language modeling (MLM) objective. Taking a sentence, the model randomly masks 15% of the words in the input then run the entire masked sentence through the model and has to predict the masked words. This is different from traditional recurrent neural networks (RNNs) that usually see the words one after the other, or from autoregressive models like GPT which internally mask the future tokens. It allows the model to learn a bidirectional representation of the sentence.

This way, the model learns an inner representation of 100 languages that can then be used to extract features useful for downstream tasks: if you have a dataset of labeled sentences for instance, you can train a standard classifier using the features produced by the XLM-RoBERTa model as inputs.

#### Limitations and bias

This model is limited by its training dataset of entity-annotated news articles from a specific span of time. This may not generalize well for all use cases in different domains.  

## Usage

You can use this model by using the AutoTokenize and AutoModelForTokenClassification class:

```python
from transformers import AutoTokenizer, AutoModelForTokenClassification 

tokenizer = AutoTokenizer.from_pretrained("julian-schelb/roberta-ner-multilingual/", add_prefix_space=True)                          
model = AutoModelForTokenClassification.from_pretrained("julian-schelb/roberta-ner-multilingual/")

text = "In December 1903 in France the Royal Swedish Academy of Sciences awarded Pierre Curie, Marie Curie, and Henri Becquerel the Nobel Prize in Physics."

inputs = tokenizer(
    text, 
    add_special_tokens=False, 
    return_tensors="pt"
)

with torch.no_grad():
    logits = model(**inputs).logits

predicted_token_class_ids = logits.argmax(-1)

# Note that tokens are classified rather then input words which means that
# there might be more predicted token classes than words.
# Multiple token classes might account for the same word
predicted_tokens_classes = [model_tuned.config.id2label[t.item()] for t in predicted_token_class_ids[0]]
predicted_tokens_classes
```

### BibTeX entry and citation info

```bibtex
TBD
```