File size: 7,629 Bytes
99d8e91
 
 
 
 
 
 
 
5b5b740
99d8e91
 
 
 
 
870b7ef
 
 
99d8e91
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
870b7ef
99d8e91
 
 
 
 
7a384d2
99d8e91
 
5b5b740
99d8e91
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5b5b740
 
99d8e91
 
7a384d2
99d8e91
 
 
7a384d2
 
 
99d8e91
7a384d2
 
99d8e91
7a384d2
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5b5b740
7a384d2
99d8e91
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
5b5b740
99d8e91
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
7a384d2
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
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
---
datasets:
- tner/tweetner7
metrics:
- f1
- precision
- recall
model-index:
- name: tner/twitter-roberta-base-2019-90m-tweetner7-all
  results:
  - task:
      name: Token Classification
      type: token-classification
    dataset:
      name: tner/tweetner7
      type: tner/tweetner7
      args: tner/tweetner7
    metrics:
    - name: F1 (test_2021)
      type: f1
      value: 0.6567966159826227
    - name: Precision (test_2021)
      type: precision
      value: 0.6494460773230839
    - name: Recall (test_2021)
      type: recall
      value: 0.6643154486586494
    - name: Macro F1 (test_2021)
      type: f1_macro
      value: 0.6099755599654287
    - name: Macro Precision (test_2021)
      type: precision_macro
      value: 0.602661693428744
    - name: Macro Recall (test_2021)
      type: recall_macro
      value: 0.6189811354202427
    - name: Entity Span F1 (test_2021)
      type: f1_entity_span
      value: 0.7888869833647745
    - name: Entity Span Precision (test_2020)
      type: precision_entity_span
      value: 0.7800135654533122
    - name: Entity Span Recall (test_2021)
      type: recall_entity_span
      value: 0.7979646120041632
    - name: F1 (test_2020)
      type: f1
      value: 0.6545553145336225
    - name: Precision (test_2020)
      type: precision
      value: 0.6854060193072118
    - name: Recall (test_2020)
      type: recall
      value: 0.6263622210690192
    - name: Macro F1 (test_2020)
      type: f1_macro
      value: 0.6121643911579755
    - name: Macro Precision (test_2020)
      type: precision_macro
      value: 0.6403532739362632
    - name: Macro Recall (test_2020)
      type: recall_macro
      value: 0.5898647290448411
    - name: Entity Span F1 (test_2020)
      type: f1_entity_span
      value: 0.7643070246813126
    - name: Entity Span Precision (test_2020)
      type: precision_entity_span
      value: 0.8005681818181818
    - name: Entity Span Recall (test_2020)
      type: recall_entity_span
      value: 0.7311883757135443

pipeline_tag: token-classification
widget:
- text: "Get the all-analog Classic Vinyl Edition of `Takin' Off` Album from {@herbiehancock@} via {@bluenoterecords@} link below: {{URL}}"
  example_title: "NER Example 1"
---
# tner/twitter-roberta-base-2019-90m-tweetner7-all

This model is a fine-tuned version of [cardiffnlp/twitter-roberta-base-2019-90m](https://huggingface.co/cardiffnlp/twitter-roberta-base-2019-90m) on the 
[tner/tweetner7](https://huggingface.co/datasets/tner/tweetner7) dataset (`train_all` split).
Model fine-tuning is done via [T-NER](https://github.com/asahi417/tner)'s hyper-parameter search (see the repository
for more detail). It achieves the following results on the test set of 2021:
- F1 (micro): 0.6567966159826227
- Precision (micro): 0.6494460773230839
- Recall (micro): 0.6643154486586494
- F1 (macro): 0.6099755599654287
- Precision (macro): 0.602661693428744
- Recall (macro): 0.6189811354202427



The per-entity breakdown of the F1 score on the test set are below:
- corporation: 0.5087071240105541
- creative_work: 0.4729907773386035
- event: 0.48405253283302063
- group: 0.6147885050048434
- location: 0.679419525065963
- person: 0.83927591881514
- product: 0.6705945366898768 

For F1 scores, the confidence interval is obtained by bootstrap as below:
- F1 (micro): 
    - 90%: [0.648368394653773, 0.6664006471768674]
    - 95%: [0.646545111092117, 0.6680503208004025] 
- F1 (macro): 
    - 90%: [0.648368394653773, 0.6664006471768674]
    - 95%: [0.646545111092117, 0.6680503208004025] 

Full evaluation can be found at [metric file of NER](https://huggingface.co/tner/twitter-roberta-base-2019-90m-tweetner7-all/raw/main/eval/metric.json) 
and [metric file of entity span](https://huggingface.co/tner/twitter-roberta-base-2019-90m-tweetner7-all/raw/main/eval/metric_span.json).

### Usage
This model can be used through the [tner library](https://github.com/asahi417/tner). Install the library via pip.   
```shell
pip install tner
```
[TweetNER7](https://huggingface.co/datasets/tner/tweetner7) pre-processed tweets where the account name and URLs are 
converted into special formats (see the dataset page for more detail), so we process tweets accordingly and then run the model prediction as below.  

```python
import re
from urlextract import URLExtract
from tner import TransformersNER

extractor = URLExtract()

def format_tweet(tweet):
    # mask web urls
    urls = extractor.find_urls(tweet)
    for url in urls:
        tweet = tweet.replace(url, "{{URL}}")
    # format twitter account
    tweet = re.sub(r"\b(\s*)(@[\S]+)\b", r'\1{\2@}', tweet)
    return tweet


text = "Get the all-analog Classic Vinyl Edition of `Takin' Off` Album from @herbiehancock via @bluenoterecords link below: http://bluenote.lnk.to/AlbumOfTheWeek"
text_format = format_tweet(text)
model = TransformersNER("tner/twitter-roberta-base-2019-90m-tweetner7-all")
model.predict([text_format])
```
It can be used via transformers library but it is not recommended as CRF layer is not supported at the moment.

### Training hyperparameters

The following hyperparameters were used during training:
 - dataset: ['tner/tweetner7']
 - dataset_split: train_all
 - dataset_name: None
 - local_dataset: None
 - model: cardiffnlp/twitter-roberta-base-2019-90m
 - crf: True
 - max_length: 128
 - epoch: 30
 - batch_size: 32
 - lr: 1e-05
 - random_seed: 0
 - gradient_accumulation_steps: 1
 - weight_decay: 1e-07
 - lr_warmup_step_ratio: 0.15
 - max_grad_norm: 1

The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/tner/twitter-roberta-base-2019-90m-tweetner7-all/raw/main/trainer_config.json).

### Reference
If you use any resource from T-NER, please consider to cite our [paper](https://aclanthology.org/2021.eacl-demos.7/).

```

@inproceedings{ushio-camacho-collados-2021-ner,
    title = "{T}-{NER}: An All-Round Python Library for Transformer-based Named Entity Recognition",
    author = "Ushio, Asahi  and
      Camacho-Collados, Jose",
    booktitle = "Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations",
    month = apr,
    year = "2021",
    address = "Online",
    publisher = "Association for Computational Linguistics",
    url = "https://aclanthology.org/2021.eacl-demos.7",
    doi = "10.18653/v1/2021.eacl-demos.7",
    pages = "53--62",
    abstract = "Language model (LM) pretraining has led to consistent improvements in many NLP downstream tasks, including named entity recognition (NER). In this paper, we present T-NER (Transformer-based Named Entity Recognition), a Python library for NER LM finetuning. In addition to its practical utility, T-NER facilitates the study and investigation of the cross-domain and cross-lingual generalization ability of LMs finetuned on NER. Our library also provides a web app where users can get model predictions interactively for arbitrary text, which facilitates qualitative model evaluation for non-expert programmers. We show the potential of the library by compiling nine public NER datasets into a unified format and evaluating the cross-domain and cross- lingual performance across the datasets. The results from our initial experiments show that in-domain performance is generally competitive across datasets. However, cross-domain generalization is challenging even with a large pretrained LM, which has nevertheless capacity to learn domain-specific features if fine- tuned on a combined dataset. To facilitate future research, we also release all our LM checkpoints via the Hugging Face model hub.",
}

```