What is the text preprocess step for dbmdz/bert-base-turkish-cased?

[daikankan]

there are some preprocess step for text before tokenize, such as https://github.com/aub-mind/arabert, but I haven't found the text preprocess step for dbmdz/bert-base-turkish-cased. Is there any thing we should do before tokenize for this model ?

[stefan-it]

Hey @daikankan ,

thanks for your interest! There were no preprocessing steps done, e.g. the vocab for the Turkish models were trained with the Tokenizers BertWordPieceTokenizer function - see an example here :)