File size: 2,421 Bytes
d0ddb4b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4e95be5
d0ddb4b
 
 
 
 
 
 
 
 
4e95be5
5e745ec
d0ddb4b
 
 
4e95be5
 
d0ddb4b
 
 
 
 
 
 
 
 
 
 
 
 
4e95be5
d0ddb4b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
license: mit
base_model: roberta-base
tags:
- generated_from_trainer
metrics:
- accuracy
- precision
- recall
- f1
model-index:
- name: irony_en_United_Kingdom
  results: []
---

# irony_en_United_Kingdom

This model is a fine-tuned version of [roberta-base](https://huggingface.co/roberta-base) on part of the [EPIC](https://huggingface.co/datasets/Multilingual-Perspectivist-NLU/EPIC) dataset.
It achieves the following results on the evaluation set:
- Loss: 0.0023
- Accuracy: 0.6833
- Precision: 0.4764
- Recall: 0.7339
- F1: 0.5778

## Model description

The model is trained considering the annotation of annotators from the United Kingdom only. 
The annotations from these annotators are aggregated using majority voting and then used to train the model.


## Training and evaluation data
The model has been trained on the annotation from annotators from the United Kingdom from the [EPIC](https://huggingface.co/datasets/Multilingual-Perspectivist-NLU/EPIC) dataset.
The data has been randomly split in a train and a validation set.

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 5e-06
- train_batch_size: 16
- eval_batch_size: 64
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: linear
- num_epochs: 20
- early stopping (patience: 2)

### Training results

| Training Loss | Epoch | Step | Validation Loss | Accuracy | Precision | Recall | F1     |
|:-------------:|:-----:|:----:|:---------------:|:--------:|:---------:|:------:|:------:|
| 0.0044        | 1.0   | 79   | 0.0042          | 0.5310   | 0.3612    | 0.7661 | 0.4910 |
| 0.0043        | 2.0   | 158  | 0.0040          | 0.6595   | 0.4417    | 0.5806 | 0.5017 |
| 0.0039        | 3.0   | 237  | 0.0034          | 0.6310   | 0.4188    | 0.6452 | 0.5079 |
| 0.0033        | 4.0   | 316  | 0.0027          | 0.7286   | 0.5352    | 0.6129 | 0.5714 |
| 0.0022        | 5.0   | 395  | 0.0024          | 0.5833   | 0.4066    | 0.8952 | 0.5592 |
| 0.0015        | 6.0   | 474  | 0.0022          | 0.7357   | 0.5474    | 0.6048 | 0.5747 |
| 0.001         | 7.0   | 553  | 0.0022          | 0.7262   | 0.5302    | 0.6371 | 0.5788 |
| 0.0005        | 8.0   | 632  | 0.0023          | 0.6833   | 0.4764    | 0.7339 | 0.5778 |


### Framework versions

- Transformers 4.34.1
- Pytorch 2.0.1+cu117
- Datasets 2.14.5
- Tokenizers 0.14.1