BenjaminOcampo
commited on
Commit
•
4304b2d
1
Parent(s):
159a0cc
Upload README.md with huggingface_hub
Browse files
README.md
ADDED
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
base_model: BenjaminOcampo/model-bert__trained-in-ishate__seed-0
|
3 |
+
datasets:
|
4 |
+
- ISHate
|
5 |
+
language:
|
6 |
+
- en
|
7 |
+
library_name: transformers
|
8 |
+
license: bsl-1.0
|
9 |
+
metrics:
|
10 |
+
- f1
|
11 |
+
- accuracy
|
12 |
+
tags:
|
13 |
+
- hate-speech-detection
|
14 |
+
- implicit-hate-speech
|
15 |
+
---
|
16 |
+
|
17 |
+
This model card documents the demo paper "PEACE: Providing Explanations and
|
18 |
+
Analysis for Combating Hate Expressions" accepted at the 27th European
|
19 |
+
Conference on Artificial Intelligence: https://www.ecai2024.eu/calls/demos.
|
20 |
+
|
21 |
+
# The Model
|
22 |
+
This model is a hate speech detector fine-tuned specifically for detecting
|
23 |
+
implicit hate speech. It is based on the paper "PEACE: Providing Explanations
|
24 |
+
and Analysis for Combating Hate Expressions" by Greta Damo, Nicolás Benjamín
|
25 |
+
Ocampo, Elena Cabrio, and Serena Villata, presented at the 27th European
|
26 |
+
Conference on Artificial Intelligence.
|
27 |
+
|
28 |
+
# Training Parameters and Experimental Info
|
29 |
+
The model was trained using the ISHate dataset, focusing on implicit data.
|
30 |
+
Training parameters included:
|
31 |
+
- Batch size: 32
|
32 |
+
- Weight decay: 0.01
|
33 |
+
- Epochs: 4
|
34 |
+
- Learning rate: 2e-5
|
35 |
+
|
36 |
+
For detailed information on the training process, please refer to the [model's
|
37 |
+
paper](https://aclanthology.org/2023.findings-emnlp.441/).
|
38 |
+
|
39 |
+
# Datasets
|
40 |
+
The model was trained on the [ISHate dataset](https://huggingface.co/datasets/BenjaminOcampo/ISHate), specifically
|
41 |
+
the training part of the dataset which focuses on implicit hate speech.
|
42 |
+
|
43 |
+
# Evaluation Results
|
44 |
+
The model's performance was evaluated using standard metrics, including F1 score
|
45 |
+
and accuracy. For comprehensive evaluation results, refer to the linked paper.
|
46 |
+
|
47 |
+
Authors:
|
48 |
+
- [Greta Damo](https://grexit-d.github.io/damo.greta.github.io/)
|
49 |
+
- [Nicolás Benjamín Ocampo](https://www.nicolasbenjaminocampo.com/)
|
50 |
+
- [Elena Cabrio](https://www-sop.inria.fr/members/Elena.Cabrio/)
|
51 |
+
- [Serena Villata](https://webusers.i3s.unice.fr/~villata/Home.html)
|