File size: 9,050 Bytes
eaf26f3
 
 
 
 
 
 
 
 
 
 
 
 
 
6b40940
eaf26f3
 
3f770c0
 
 
 
eaf26f3
 
 
6a4e40d
eaf26f3
6a4e40d
eaf26f3
6a4e40d
eaf26f3
6a4e40d
eaf26f3
6a4e40d
eaf26f3
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
library_name: transformers
tags:
- generated_from_trainer
model-index:
- name: gbert-base-defakts-fake-binary
  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. -->

# gbert-base-defakts-fake-binary

This Model is finetuned for sequence classification (binary fake-news classification task) on the german DeFaktS-Dataset.
It achieves the following results on the evaluation set:
- Loss: 0.3441
- Accuracy: 0.8526
- F1: 0.8413
- Precision: 0.8545
- Recall: 0.8337

## Model description

This Model is finetuned for sequence classification

### Dataset

Trained on the DeFactS dataset https://github.com/caisa-lab/DeFaktS-Dataset-Disinformaton-Detection, feature catposfake/catneutral to detect fake news

## Intended uses & limitations

Fake news classification

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 5e-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: 6

### Training results

| Training Loss | Epoch  | Step | Validation Loss | Accuracy@de | F1@de  | Precision@de | Recall@de | Loss@de |
|:-------------:|:------:|:----:|:---------------:|:-----------:|:------:|:------------:|:---------:|:-------:|
| 0.6236        | 0.0888 | 50   | 0.5162          | 0.7226      | 0.7219 | 0.7377       | 0.7460    | 0.5162  |
| 0.4355        | 0.1776 | 100  | 0.4000          | 0.8201      | 0.8112 | 0.8116       | 0.8107    | 0.4000  |
| 0.3937        | 0.2664 | 150  | 0.3916          | 0.8316      | 0.8181 | 0.8323       | 0.8106    | 0.3917  |
| 0.409         | 0.3552 | 200  | 0.3726          | 0.8331      | 0.8216 | 0.8299       | 0.8163    | 0.3727  |
| 0.3579        | 0.4440 | 250  | 0.3572          | 0.8386      | 0.8285 | 0.8339       | 0.8246    | 0.3573  |
| 0.366         | 0.5329 | 300  | 0.3519          | 0.8456      | 0.8370 | 0.8396       | 0.8349    | 0.3521  |
| 0.3735        | 0.6217 | 350  | 0.4180          | 0.7991      | 0.7976 | 0.8038       | 0.8179    | 0.4180  |
| 0.3849        | 0.7105 | 400  | 0.3605          | 0.8386      | 0.8198 | 0.8592       | 0.8064    | 0.3607  |
| 0.3752        | 0.7993 | 450  | 0.3379          | 0.8581      | 0.8532 | 0.8499       | 0.8586    | 0.3379  |
| 0.3547        | 0.8881 | 500  | 0.3441          | 0.8526      | 0.8413 | 0.8545       | 0.8337    | 0.3442  |
| 0.344         | 0.9769 | 550  | 0.3439          | 0.8461      | 0.8320 | 0.8540       | 0.8218    | 0.3440  |
| 0.2853        | 1.0657 | 600  | 0.3525          | 0.8446      | 0.8386 | 0.8361       | 0.8421    | 0.3526  |
| 0.2505        | 1.1545 | 650  | 0.3427          | 0.8626      | 0.8512 | 0.8686       | 0.8419    | 0.3429  |
| 0.2358        | 1.2433 | 700  | 0.3537          | 0.8626      | 0.8556 | 0.8565       | 0.8547    | 0.3537  |
| 0.257         | 1.3321 | 750  | 0.3258          | 0.8621      | 0.8559 | 0.8548       | 0.8572    | 0.3259  |
| 0.2369        | 1.4210 | 800  | 0.3810          | 0.8551      | 0.8505 | 0.8469       | 0.8571    | 0.3811  |
| 0.248         | 1.5098 | 850  | 0.3576          | 0.8721      | 0.8641 | 0.8701       | 0.8596    | 0.3578  |
| 0.2612        | 1.5986 | 900  | 0.3273          | 0.8686      | 0.8583 | 0.8729       | 0.8500    | 0.3275  |
| 0.2532        | 1.6874 | 950  | 0.3235          | 0.8636      | 0.8567 | 0.8575       | 0.8560    | 0.3236  |
| 0.2225        | 1.7762 | 1000 | 0.3513          | 0.8666      | 0.8567 | 0.8690       | 0.8492    | 0.3515  |
| 0.2497        | 1.8650 | 1050 | 0.3497          | 0.8711      | 0.8625 | 0.8705       | 0.8570    | 0.3498  |
| 0.2291        | 1.9538 | 1100 | 0.3395          | 0.8761      | 0.8685 | 0.8740       | 0.8643    | 0.3396  |
| 0.1675        | 2.0426 | 1150 | 0.3944          | 0.8671      | 0.8583 | 0.8660       | 0.8530    | 0.3946  |
| 0.1182        | 2.1314 | 1200 | 0.4743          | 0.8626      | 0.8532 | 0.8621       | 0.8473    | 0.4746  |
| 0.1453        | 2.2202 | 1250 | 0.4977          | 0.8646      | 0.8531 | 0.8719       | 0.8433    | 0.4981  |
| 0.1288        | 2.3091 | 1300 | 0.4345          | 0.8696      | 0.8631 | 0.8637       | 0.8625    | 0.4347  |
| 0.1399        | 2.3979 | 1350 | 0.4128          | 0.8751      | 0.8665 | 0.8758       | 0.8603    | 0.4132  |
| 0.1384        | 2.4867 | 1400 | 0.3688          | 0.8776      | 0.8713 | 0.8723       | 0.8704    | 0.3690  |
| 0.1292        | 2.5755 | 1450 | 0.4154          | 0.8781      | 0.8710 | 0.8749       | 0.8679    | 0.4157  |
| 0.112         | 2.6643 | 1500 | 0.4399          | 0.8661      | 0.8592 | 0.8603       | 0.8583    | 0.4401  |
| 0.108         | 2.7531 | 1550 | 0.4439          | 0.8731      | 0.8659 | 0.8692       | 0.8631    | 0.4442  |
| 0.1153        | 2.8419 | 1600 | 0.4476          | 0.8676      | 0.8590 | 0.8662       | 0.8539    | 0.4479  |
| 0.1228        | 2.9307 | 1650 | 0.4933          | 0.8691      | 0.8589 | 0.8733       | 0.8506    | 0.4936  |
| 0.0955        | 3.0195 | 1700 | 0.5272          | 0.8806      | 0.8726 | 0.8810       | 0.8668    | 0.5275  |
| 0.0502        | 3.1083 | 1750 | 0.6531          | 0.8661      | 0.8554 | 0.8708       | 0.8468    | 0.6537  |
| 0.0604        | 3.1972 | 1800 | 0.6515          | 0.8721      | 0.8635 | 0.8719       | 0.8578    | 0.6520  |
| 0.068         | 3.2860 | 1850 | 0.6422          | 0.8756      | 0.8656 | 0.8820       | 0.8564    | 0.6427  |
| 0.0505        | 3.3748 | 1900 | 0.6262          | 0.8681      | 0.8606 | 0.8639       | 0.8579    | 0.6266  |
| 0.065         | 3.4636 | 1950 | 0.6342          | 0.8681      | 0.8614 | 0.8623       | 0.8606    | 0.6345  |
| 0.0645        | 3.5524 | 2000 | 0.6472          | 0.8696      | 0.8623 | 0.8653       | 0.8598    | 0.6476  |
| 0.0951        | 3.6412 | 2050 | 0.6048          | 0.8661      | 0.8576 | 0.8641       | 0.8529    | 0.6051  |
| 0.0336        | 3.7300 | 2100 | 0.6603          | 0.8736      | 0.8661 | 0.8705       | 0.8626    | 0.6608  |
| 0.0551        | 3.8188 | 2150 | 0.6932          | 0.8716      | 0.8638 | 0.8689       | 0.8599    | 0.6938  |
| 0.0595        | 3.9076 | 2200 | 0.6379          | 0.8756      | 0.8687 | 0.8715       | 0.8663    | 0.6384  |
| 0.0681        | 3.9964 | 2250 | 0.6327          | 0.8751      | 0.8661 | 0.8774       | 0.8589    | 0.6332  |
| 0.0273        | 4.0853 | 2300 | 0.6414          | 0.8731      | 0.8656 | 0.8701       | 0.8620    | 0.6419  |
| 0.0261        | 4.1741 | 2350 | 0.6590          | 0.8761      | 0.8699 | 0.8705       | 0.8692    | 0.6594  |
| 0.0173        | 4.2629 | 2400 | 0.7341          | 0.8776      | 0.8703 | 0.8750       | 0.8666    | 0.7347  |
| 0.024         | 4.3517 | 2450 | 0.7647          | 0.8706      | 0.8639 | 0.8652       | 0.8626    | 0.7651  |
| 0.0324        | 4.4405 | 2500 | 0.7651          | 0.8741      | 0.8657 | 0.8738       | 0.8601    | 0.7658  |
| 0.0144        | 4.5293 | 2550 | 0.7918          | 0.8691      | 0.8599 | 0.8698       | 0.8535    | 0.7925  |
| 0.0402        | 4.6181 | 2600 | 0.7661          | 0.8691      | 0.8604 | 0.8684       | 0.8549    | 0.7667  |
| 0.0311        | 4.7069 | 2650 | 0.7688          | 0.8706      | 0.8619 | 0.8701       | 0.8564    | 0.7694  |
| 0.0127        | 4.7957 | 2700 | 0.7880          | 0.8691      | 0.8607 | 0.8675       | 0.8558    | 0.7886  |
| 0.03          | 4.8845 | 2750 | 0.7616          | 0.8736      | 0.8672 | 0.8680       | 0.8665    | 0.7620  |
| 0.0431        | 4.9734 | 2800 | 0.7842          | 0.8716      | 0.8631 | 0.8710       | 0.8576    | 0.7848  |
| 0.0243        | 5.0622 | 2850 | 0.7620          | 0.8701      | 0.8612 | 0.8704       | 0.8550    | 0.7626  |
| 0.0135        | 5.1510 | 2900 | 0.7799          | 0.8711      | 0.8624 | 0.8710       | 0.8565    | 0.7805  |
| 0.0177        | 5.2398 | 2950 | 0.7644          | 0.8746      | 0.8672 | 0.8716       | 0.8637    | 0.7649  |
| 0.0126        | 5.3286 | 3000 | 0.7826          | 0.8736      | 0.8656 | 0.8721       | 0.8608    | 0.7832  |
| 0.011         | 5.4174 | 3050 | 0.7951          | 0.8761      | 0.8681 | 0.8751       | 0.8631    | 0.7957  |
| 0.0214        | 5.5062 | 3100 | 0.7953          | 0.8741      | 0.8657 | 0.8738       | 0.8601    | 0.7960  |
| 0.0099        | 5.5950 | 3150 | 0.7855          | 0.8746      | 0.8667 | 0.8729       | 0.8621    | 0.7861  |
| 0.0193        | 5.6838 | 3200 | 0.7967          | 0.8746      | 0.8661 | 0.8747       | 0.8603    | 0.7974  |
| 0.0171        | 5.7726 | 3250 | 0.7956          | 0.8751      | 0.8669 | 0.8744       | 0.8616    | 0.7962  |
| 0.013         | 5.8615 | 3300 | 0.7972          | 0.8741      | 0.8658 | 0.8736       | 0.8604    | 0.7978  |
| 0.0176        | 5.9503 | 3350 | 0.8003          | 0.8736      | 0.8651 | 0.8734       | 0.8595    | 0.8009  |


### Framework versions

- Transformers 4.45.2
- Pytorch 2.3.1+cu121
- Tokenizers 0.20.3