File size: 5,268 Bytes
15a607c
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
---
library_name: transformers
license: apache-2.0
base_model: distilbert/distilroberta-base
tags:
- generated_from_trainer
- sentiment_analysis
model-index:
- name: augmented-go-emotions-plus-other-datasets-fine-tuned-distilroberta
  results: []
datasets:
- google-research-datasets/go_emotions
language:
- en
metrics:
- f1
- precision
- recall
---

<!-- 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. -->

# augmented-go-emotions-plus-other-datasets-fine-tuned-distilroberta

This model is a fine-tuned version of [distilbert/distilroberta-base](https://huggingface.co/distilbert/distilroberta-base) on the these datasets:
   - [GoEmotions](https://github.com/google-research/google-research/tree/master/goemotions)
   - [sem_eval_2018_task_1 (English)](https://huggingface.co/datasets/SemEvalWorkshop/sem_eval_2018_task_1)
   - [Emotion Detection from Text - Pashupati Gupta](https://www.kaggle.com/datasets/pashupatigupta/emotion-detection-from-text/data)
   - [Emotions dataset for NLP - praveengovi](https://www.kaggle.com/datasets/praveengovi/emotions-dataset-for-nlp/data)
It has also been data augmented using TextAttack.

It achieves the following results on the evaluation set:
- Loss: 0.0731
- Micro Precision: 0.7189
- Micro Recall: 0.5774
- Micro F1: 0.6404
- Macro Precision: 0.6049
- Macro Recall: 0.4433
- Macro F1: 0.4898
- Weighted Precision: 0.7004
- Weighted Recall: 0.5774
- Weighted F1: 0.6243
- Hamming Loss: 0.0276


### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 5e-05
- train_batch_size: 8
- eval_batch_size: 8
- seed: 42
- optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
- lr_scheduler_type: linear
- num_epochs: 3.0

### Training results

| Training Loss | Epoch | Step  | Validation Loss | Micro Precision | Micro Recall | Micro F1 | Macro Precision | Macro Recall | Macro F1 | Weighted Precision | Weighted Recall | Weighted F1 | Hamming Loss |
|:-------------:|:-----:|:-----:|:---------------:|:---------------:|:------------:|:--------:|:---------------:|:------------:|:--------:|:------------------:|:---------------:|:-----------:|:------------:|
| No log        | 1.0   | 11118 | 0.0765          | 0.7647          | 0.5046       | 0.6080   | 0.6047          | 0.3580       | 0.4127   | 0.7321             | 0.5046          | 0.5764      | 0.0277       |
| No log        | 2.0   | 22236 | 0.0733          | 0.7309          | 0.5344       | 0.6174   | 0.5791          | 0.4162       | 0.4611   | 0.7105             | 0.5344          | 0.5923      | 0.0282       |
| No log        | 3.0   | 33354 | 0.0731          | 0.7189          | 0.5774       | 0.6404   | 0.6049          | 0.4433       | 0.4898   | 0.7004             | 0.5774          | 0.6243      | 0.0276       |


### Test results
Threshold = 0.5
| Label            | Precision | Recall | F1-Score | Support |
|------------------|-----------|--------|----------|---------|
| admiration       | 0.65      | 0.70   | 0.67     | 504     |
| amusement        | 0.72      | 0.88   | 0.79     | 264     |
| anger            | 0.79      | 0.69   | 0.73     | 1585    |
| annoyance        | 0.45      | 0.12   | 0.19     | 320     |
| approval         | 0.63      | 0.27   | 0.38     | 351     |
| caring           | 0.44      | 0.36   | 0.40     | 135     |
| confusion        | 0.44      | 0.39   | 0.41     | 153     |
| curiosity        | 0.52      | 0.36   | 0.43     | 284     |
| desire           | 0.50      | 0.37   | 0.43     | 83      |
| disappointment   | 0.35      | 0.19   | 0.25     | 151     |
| disapproval      | 0.49      | 0.31   | 0.38     | 267     |
| disgust          | 0.72      | 0.62   | 0.66     | 1222    |
| embarrassment    | 0.68      | 0.35   | 0.46     | 37      |
| excitement       | 0.46      | 0.43   | 0.44     | 103     |
| fear             | 0.82      | 0.73   | 0.77     | 787     |
| gratitude        | 0.93      | 0.89   | 0.91     | 352     |
| grief            | 0.00      | 0.00   | 0.00     | 6       |
| joy              | 0.85      | 0.78   | 0.81     | 2298    |
| love             | 0.70      | 0.60   | 0.65     | 1305    |
| nervousness      | 0.44      | 0.17   | 0.25     | 23      |
| optimism         | 0.70      | 0.56   | 0.62     | 1329    |
| pride            | 0.00      | 0.00   | 0.00     | 16      |
| realization      | 0.36      | 0.17   | 0.23     | 145     |
| relief           | 0.28      | 0.22   | 0.24     | 160     |
| remorse          | 0.59      | 0.80   | 0.68     | 56      |
| sadness          | 0.78      | 0.66   | 0.71     | 2212    |
| surprise         | 0.63      | 0.29   | 0.40     | 572     |
| neutral          | 0.70      | 0.52   | 0.60     | 2668    |
| **Micro Avg**    | 0.73      | 0.59   | 0.65     | 17388   |
| **Macro Avg**    | 0.56      | 0.44   | 0.48     | 17388   |
| **Weighted Avg** | 0.72      | 0.59   | 0.64     | 17388   |
| **Samples Avg**  | 0.63      | 0.60   | 0.60     | 17388   |


### Framework versions

- Transformers 4.47.0
- Pytorch 2.3.1+cu121
- Datasets 2.20.0
- Tokenizers 0.21.0