Datasets:

Languages:
English
ArXiv:
License:
File size: 4,116 Bytes
6c8a3a5
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import os
from typing import Dict


def get_readme(model_name: str,
               metric: Dict,
               metric_span: Dict,
               config: Dict):
    language_model = config['model']
    dataset = None
    dataset_alias = "custom"
    if config["dataset"] is not None:
        dataset = sorted([i for i in config["dataset"]])
        dataset_alias = ','.join(dataset)
    config_text = "\n".join([f" - {k}: {v}" for k, v in config.items()])
    ci_micro = '\n'.join([f'    - {k}%: {v}' for k, v in metric["micro/f1_ci"].items()])
    ci_macro = '\n'.join([f'    - {k}%: {v}' for k, v in metric["micro/f1_ci"].items()])
    per_entity_metric = '\n'.join([f'- {k}: {v["f1"]}' for k, v in metric['per_entity_metric'].items()])
    if dataset is None:
        dataset_link = 'custom'
    else:
        dataset = [dataset] if type(dataset) is str else dataset
        dataset_link = ','.join([f"[{d}](https://huggingface.co/datasets/{d})" for d in dataset])
    return f"""---
datasets:
- {dataset_alias}
metrics:
- f1
- precision
- recall
model-index:
- name: {model_name}
  results:
  - task:
      name: Token Classification
      type: token-classification
    dataset:
      name: {dataset_alias}
      type: {dataset_alias}
      args: {dataset_alias}
    metrics:
    - name: F1
      type: f1
      value: {metric['micro/f1']}
    - name: Precision
      type: precision
      value: {metric['micro/precision']}
    - name: Recall
      type: recall
      value: {metric['micro/recall']}
    - name: F1 (macro)
      type: f1_macro
      value: {metric['macro/f1']}
    - name: Precision (macro)
      type: precision_macro
      value: {metric['macro/precision']}
    - name: Recall (macro)
      type: recall_macro
      value: {metric['macro/recall']}
    - name: F1 (entity span)
      type: f1_entity_span
      value: {metric_span['micro/f1']}
    - name: Precision (entity span)
      type: precision_entity_span
      value: {metric_span['micro/precision']}
    - name: Recall (entity span)
      type: recall_entity_span
      value: {metric_span['micro/recall']}

pipeline_tag: token-classification
widget:
- text: "Jacob Collier is a Grammy awarded artist from England."
  example_title: "NER Example 1"
---
# {model_name}

This model is a fine-tuned version of [{language_model}](https://huggingface.co/{language_model}) on the 
{dataset_link} dataset.
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:
- F1 (micro): {metric['micro/f1']}
- Precision (micro): {metric['micro/precision']}
- Recall (micro): {metric['micro/recall']}
- F1 (macro): {metric['macro/f1']}
- Precision (macro): {metric['macro/precision']}
- Recall (macro): {metric['macro/recall']}

The per-entity breakdown of the F1 score on the test set are below:
{per_entity_metric} 

For F1 scores, the confidence interval is obtained by bootstrap as below:
- F1 (micro): 
{ci_micro} 
- F1 (macro): 
{ci_macro} 

Full evaluation can be found at [metric file of NER](https://huggingface.co/{model_name}/raw/main/eval/metric.json) 
and [metric file of entity span](https://huggingface.co/{model_name}/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
```
and activate model as below.
```python
from tner import TransformersNER
model = TransformersNER("{model_name}")
model.predict(["Jacob Collier is a Grammy awarded English artist from London"])
```
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:
{config_text}

The full configuration can be found at [fine-tuning parameter file](https://huggingface.co/{model_name}/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/).

```
{bib}
```
"""