File size: 3,832 Bytes
48ceca1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2d19cbc
 
 
 
 
 
 
48ceca1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
2d19cbc
 
 
 
 
 
 
48ceca1
 
 
 
 
 
 
 
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
---
license: mit
base_model: dathi103/gbert-job-extended
tags:
- generated_from_trainer
model-index:
- name: gerskill-gbert-job-extended
  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. -->

# gerskill-gbert-job-extended

This model is a fine-tuned version of [dathi103/gbert-job-extended](https://huggingface.co/dathi103/gbert-job-extended) on an unknown dataset.
It achieves the following results on the evaluation set:
- Loss: 0.1217
- Hard: {'precision': 0.7340153452685422, 'recall': 0.790633608815427, 'f1': 0.7612732095490715, 'number': 363}
- Soft: {'precision': 0.6911764705882353, 'recall': 0.7121212121212122, 'f1': 0.7014925373134329, 'number': 66}
- Overall Precision: 0.7277
- Overall Recall: 0.7786
- Overall F1: 0.7523
- Overall Accuracy: 0.9661

## 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: 2e-05
- train_batch_size: 32
- eval_batch_size: 32
- 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 | Hard                                                                                                     | Soft                                                                                                    | Overall Precision | Overall Recall | Overall F1 | Overall Accuracy |
|:-------------:|:-----:|:----:|:---------------:|:--------------------------------------------------------------------------------------------------------:|:-------------------------------------------------------------------------------------------------------:|:-----------------:|:--------------:|:----------:|:----------------:|
| No log        | 1.0   | 178  | 0.1108          | {'precision': 0.6256038647342995, 'recall': 0.7134986225895317, 'f1': 0.6666666666666667, 'number': 363} | {'precision': 0.5606060606060606, 'recall': 0.5606060606060606, 'f1': 0.5606060606060606, 'number': 66} | 0.6167            | 0.6900         | 0.6513     | 0.9593           |
| No log        | 2.0   | 356  | 0.1027          | {'precision': 0.6860759493670886, 'recall': 0.7465564738292011, 'f1': 0.7150395778364115, 'number': 363} | {'precision': 0.7096774193548387, 'recall': 0.6666666666666666, 'f1': 0.6875, 'number': 66}             | 0.6893            | 0.7343         | 0.7111     | 0.9639           |
| 0.1153        | 3.0   | 534  | 0.1085          | {'precision': 0.7085427135678392, 'recall': 0.7768595041322314, 'f1': 0.7411300919842312, 'number': 363} | {'precision': 0.6533333333333333, 'recall': 0.7424242424242424, 'f1': 0.6950354609929078, 'number': 66} | 0.6998            | 0.7716         | 0.7339     | 0.9658           |
| 0.1153        | 4.0   | 712  | 0.1163          | {'precision': 0.6987341772151898, 'recall': 0.7603305785123967, 'f1': 0.7282321899736148, 'number': 363} | {'precision': 0.7121212121212122, 'recall': 0.7121212121212122, 'f1': 0.7121212121212122, 'number': 66} | 0.7007            | 0.7529         | 0.7258     | 0.9657           |
| 0.1153        | 5.0   | 890  | 0.1217          | {'precision': 0.7340153452685422, 'recall': 0.790633608815427, 'f1': 0.7612732095490715, 'number': 363}  | {'precision': 0.6911764705882353, 'recall': 0.7121212121212122, 'f1': 0.7014925373134329, 'number': 66} | 0.7277            | 0.7786         | 0.7523     | 0.9661           |


### Framework versions

- Transformers 4.38.1
- Pytorch 2.1.2+cu121
- Datasets 2.18.0
- Tokenizers 0.15.2