File size: 5,386 Bytes
53f7621
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: cc-by-sa-4.0
tags:
- generated_from_trainer
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: legal-NER
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# legal-NER

This model is a fine-tuned version of [nlpaueb/legal-bert-base-uncased](https://huggingface.co/nlpaueb/legal-bert-base-uncased) on the None dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0068
- Accuracy: 0.9990
- Precision: 0.9931
- Recall: 0.9944
- F1: 0.9938
- Classification Report:               precision    recall  f1-score   support

         LOC       1.00      1.00      1.00      1837
        MISC       0.98      0.98      0.98       922
         ORG       1.00      0.99      0.99      1341
         PER       1.00      1.00      1.00      1842

   micro avg       0.99      0.99      0.99      5942
   macro avg       0.99      0.99      0.99      5942
weighted avg       0.99      0.99      0.99      5942


## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 16
- eval_batch_size: 16
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 5

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1     | Classification Report                                                                                                                                                                                                                                                                                                                                                                                                                              |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|:--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------:|
| 0.1501        | 1.0   | 217  | 0.0704          | 0.9810   | 0.8615    | 0.8901 | 0.8756 |               precision    recall  f1-score   support

         LOC       0.86      0.95      0.91      1837
        MISC       0.74      0.70      0.72       922
         ORG       0.80      0.82      0.81      1341
         PER       0.97      0.97      0.97      1842

   micro avg       0.86      0.89      0.88      5942
   macro avg       0.84      0.86      0.85      5942
weighted avg       0.86      0.89      0.87      5942
 |
| 0.0682        | 2.0   | 434  | 0.0266          | 0.9929   | 0.9513    | 0.9631 | 0.9572 |               precision    recall  f1-score   support

         LOC       0.98      0.98      0.98      1837
        MISC       0.88      0.91      0.89       922
         ORG       0.92      0.96      0.94      1341
         PER       0.99      0.97      0.98      1842

   micro avg       0.95      0.96      0.96      5942
   macro avg       0.94      0.96      0.95      5942
weighted avg       0.95      0.96      0.96      5942
 |
| 0.0362        | 3.0   | 651  | 0.0137          | 0.9970   | 0.9776    | 0.9850 | 0.9813 |               precision    recall  f1-score   support

         LOC       0.98      1.00      0.99      1837
        MISC       0.94      0.95      0.94       922
         ORG       0.98      0.98      0.98      1341
         PER       0.99      1.00      1.00      1842

   micro avg       0.98      0.99      0.98      5942
   macro avg       0.97      0.98      0.98      5942
weighted avg       0.98      0.99      0.98      5942
 |
| 0.0209        | 4.0   | 868  | 0.0079          | 0.9986   | 0.9894    | 0.9918 | 0.9906 |               precision    recall  f1-score   support

         LOC       0.99      1.00      1.00      1837
        MISC       0.98      0.97      0.97       922
         ORG       0.99      0.99      0.99      1341
         PER       1.00      1.00      1.00      1842

   micro avg       0.99      0.99      0.99      5942
   macro avg       0.99      0.99      0.99      5942
weighted avg       0.99      0.99      0.99      5942
 |
| 0.0143        | 5.0   | 1085 | 0.0068          | 0.9990   | 0.9931    | 0.9944 | 0.9938 |               precision    recall  f1-score   support

         LOC       1.00      1.00      1.00      1837
        MISC       0.98      0.98      0.98       922
         ORG       1.00      0.99      0.99      1341
         PER       1.00      1.00      1.00      1842

   micro avg       0.99      0.99      0.99      5942
   macro avg       0.99      0.99      0.99      5942
weighted avg       0.99      0.99      0.99      5942
 |


### Framework versions

- Transformers 4.30.2
- Pytorch 2.0.0
- Datasets 2.1.0
- Tokenizers 0.13.3