Text classification model based on EMBEDDIA/sloberta and fine-tuned on the FRENK dataset comprising of LGBT and migrant hatespeech. Only the slovenian subset of the data was used for fine-tuning and the dataset has been relabeled for binary classification (offensive or acceptable).
Fine-tuning hyperparameters
Fine-tuning was performed with simpletransformers. Beforehand a brief hyperparameter optimisation was performed and the presumed optimal hyperparameters are:
The same pipeline was run with two other transformer models and fasttext for comparison. Accuracy and macro F1 score were recorded for each of the 6 fine-tuning sessions and post festum analyzed.
model
average accuracy
average macro F1
sloberta-frenk-hate
0.7785
0.7764
EMBEDDIA/crosloengual-bert
0.7616
0.7585
xlm-roberta-base
0.686
0.6827
fasttext
0.709
0.701
From recorded accuracies and macro F1 scores p-values were also calculated:
Comparison with crosloengual-bert:
test
accuracy p-value
macro F1 p-value
Wilcoxon
0.00781
0.00781
Mann Whithney U test
0.00163
0.00108
Student t-test
0.000101
3.95e-05
Comparison with xlm-roberta-base:
test
accuracy p-value
macro F1 p-value
Wilcoxon
0.00781
0.00781
Mann Whithney U test
0.00108
0.00108
Student t-test
9.46e-11
6.94e-11
Use examples
python
1from simpletransformers.classification import ClassificationModel
2model_args ={3"num_train_epochs":6,4"learning_rate":3e-6,5"train_batch_size":69}67model = ClassificationModel(8"camembert","5roop/sloberta-frenk-hate", use_cuda=True,9 args=model_args
1011)1213predictions, logit_output = model.predict(["Silva, ti si grda in neprijazna","Naša hiša ima dimnik"])14predictions
15### Output:16### array([1, 0])
Citation
If you use the model, please cite the following paper on which the original model is based:
@article{DBLP:journals/corr/abs-1907-11692,
author = {Yinhan Liu and
Myle Ott and
Naman Goyal and
Jingfei Du and
Mandar Joshi and
Danqi Chen and
Omer Levy and
Mike Lewis and
Luke Zettlemoyer and
Veselin Stoyanov},
title = {RoBERTa: {A} Robustly Optimized {BERT} Pretraining Approach},
journal = {CoRR},
volume = {abs/1907.11692},
year = {2019},
url = {http://arxiv.org/abs/1907.11692},
archivePrefix = {arXiv},
eprint = {1907.11692},
timestamp = {Thu, 01 Aug 2019 08:59:33 +0200},
biburl = {https://dblp.org/rec/journals/corr/abs-1907-11692.bib},
bibsource = {dblp computer science bibliography, https://dblp.org}
}
and the dataset used for fine-tuning:
@misc{ljubešić2019frenk,
title={The FRENK Datasets of Socially Unacceptable Discourse in Slovene and English},
author={Nikola Ljubešić and Darja Fišer and Tomaž Erjavec},
year={2019},
eprint={1906.02045},
archivePrefix={arXiv},
primaryClass={cs.CL},
url={https://arxiv.org/abs/1906.02045}
}
