File size: 1,109 Bytes
58f55e3 edbbde9 58f55e3 a9c60aa 58f55e3 edbbde9 58f55e3 daf9d09 4c88e0e 58f55e3 daf9d09 58f55e3 daf9d09 58f55e3 4c88e0e 4b56b80 4c88e0e 4b56b80 4c88e0e 58f55e3 f0fdaf4 daf9d09 58f55e3 eb65526 58f55e3 |
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 |
---
language:
- da
tags:
- Hatespeech
- Danish
- BERT
license: mit
datasets:
- DKHate - OffensEval2020
Classes:
- Hateful
- Not Hateful
---
# DKbert-hatespeech-classification
Use this model to detect hatespeech in Danish. For details, guide and command line tool see [DK hate github](https://github.com/Guscode/DKbert-hatespeech-detection)
## Training data
Training data is from OffensEval2020 which can be found [here]( https://figshare.com/articles/dataset/Danish_Hate_Speech_Abusive_Language_data/12220805)
## Performance
The model achieves a macro F1-score of 0.78
Precision hateful: 0.77
Recall hateful: 0.49
See more on [DK hate github](https://github.com/Guscode/DKbert-hatespeech-detection)
## Training procedure
- [BOTXO Nordic Bert](https://huggingface.co/DJSammy/bert-base-danish-uncased_BotXO,ai)
- Learning rate: 1e-5,
- Batch size: 16
- Max sequence length: 128
## Project information
This model was made in collaboration between [Johan Horsmans](https://github.com/JohanHorsmans) and [Gustav Aarup Lauridsen](https://github.com/Guscode) for their Cultural Data Science Exam.
|