File size: 2,907 Bytes
b08c525
08c0f38
b08c525
 
 
52ddf8a
b08c525
 
 
 
 
6d6c977
 
 
 
 
 
b08c525
 
 
 
 
6d6c977
b08c525
 
 
 
 
 
 
6d6c977
b08c525
6d6c977
 
b08c525
6d6c977
 
b08c525
6d6c977
 
b08c525
6d6c977
b08c525
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
bc78d5b
b08c525
 
 
bc78d5b
b08c525
 
 
bc78d5b
b08c525
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: mit
tags:
- generated_from_trainer
datasets:
- tner/ontonotes5
metrics:
- precision
- recall
- f1
- accuracy
widget:
- text: 'I am Jack. I live in California and I work at Apple '
  example_title: Example 1
- text: 'Wow this book is amazing and costs only 4€ '
  example_title: Example 2
base_model: distilbert-base-cased
model-index:
- name: distilbert-finetuned-ner-ontonotes
  results:
  - task:
      type: token-classification
      name: Token Classification
    dataset:
      name: ontonotes5
      type: ontonotes5
      config: ontonotes5
      split: train
      args: ontonotes5
    metrics:
    - type: precision
      value: 0.8535359959297889
      name: Precision
    - type: recall
      value: 0.8788553467356427
      name: Recall
    - type: f1
      value: 0.8660106468785288
      name: F1
    - type: accuracy
      value: 0.9749625470373822
      name: Accuracy
---

<!-- 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. -->

# distilbert-finetuned-ner-ontonotes

This model is a fine-tuned version of [distilbert-base-cased](https://huggingface.co/distilbert-base-cased) on the ontonotes5 dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1448
- Precision: 0.8535
- Recall: 0.8789
- F1: 0.8660
- Accuracy: 0.9750

## Model description

Token classification experiment, NER, on business topics.

## Intended uses & limitations

The model can be used on token classification, in particular NER. It is fine tuned on business domain.

## Training and evaluation data

The dataset used is [ontonotes5](https://huggingface.co/datasets/tner/ontonotes5)

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 2e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 6

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Precision | Recall | F1     | Accuracy |
|:-------------:|:-----:|:-----:|:---------------:|:---------:|:------:|:------:|:--------:|
| 0.0937        | 1.0   | 7491  | 0.0998          | 0.8367    | 0.8587 | 0.8475 | 0.9731   |
| 0.0572        | 2.0   | 14982 | 0.1084          | 0.8338    | 0.8759 | 0.8543 | 0.9737   |
| 0.0403        | 3.0   | 22473 | 0.1145          | 0.8521    | 0.8707 | 0.8613 | 0.9748   |
| 0.0265        | 4.0   | 29964 | 0.1222          | 0.8535    | 0.8815 | 0.8672 | 0.9752   |
| 0.0148        | 5.0   | 37455 | 0.1365          | 0.8536    | 0.8770 | 0.8651 | 0.9747   |
| 0.0111        | 6.0   | 44946 | 0.1448          | 0.8535    | 0.8789 | 0.8660 | 0.9750   |


### Framework versions

- Transformers 4.22.1
- Pytorch 1.12.1+cu113
- Datasets 2.5.1
- Tokenizers 0.12.1