This is a simple demo setup for SloNSpell, a 🇸🇮 Slovene spelling error detection model.
You can find more about the model in the model card cjvt/SloBERTa-slo-word-spelling-annotator.
Given an input text:
1. The input is segmented into sentences and tokenized using NLTK to prepare the model input.
2. The model makes predictions on the sentence level.
The model does not work perfectly and can make mistakes, please check the output!