File size: 1,034 Bytes
56e6d8b
 
b0a4a47
 
 
56e6d8b
b0a4a47
a4c0dc4
b0a4a47
a4c0dc4
120b3da
 
a4c0dc4
120b3da
 
 
a4c0dc4
120b3da
a4c0dc4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
---
license: apache-2.0
language:
- nl
pipeline_tag: text-classification
---

Fine-tuned model for detecting instances of offensive language in Dutch tweets. The model has been trained with [DALC v2.0 ](https://github.com/tommasoc80/DALC).

Offensive language definition is inherited from SemEval 2019 OffensEval: "Posts containing any form of non-acceptable language (profanity) or a targeted offence,
which can be veiled or direct. This includes insults, threats, and posts containing profane language or swear words." ([Zampieri et al., 2019](https://aclanthology.org/N19-1144/))

The model achieves the following results on multiple test data:

- DALC held-out test set: macro F1: 79.93; F1 Offensive: 70.34
- HateCheck-NL (functional benchmark for hate speech): Accuracy: 61.40;  Accuracy non-hateful tests: 47.61 ; Accuracy hateful tests: 68.86
- OP-NL (dynamic benchmark for offensive language): macro F1: 73.56

More details on the training settings and pre-processing are available [here](https://github.com/tommasoc80/DALC)