Feature extraction

[aretii]

Is there any way to use fietje for textual feature extraction? As it is trained in a big amount of Dutch data, I would like to use for feature extraction from dutch text data.
Thank you in advance!

[BramVanroy]

I would personally instead use an encoder model or a specific embedding model to get latent features from it. Those are more intended for this task.

My suggestions:

• https://huggingface.co/DTAI-KULeuven/robbert-2022-dutch-base
• https://huggingface.co/NetherlandsForensicInstitute/robbert-2022-dutch-sentence-transformers
• https://huggingface.co/jinaai/jina-embeddings-v3

[aretii]

Thanks a lot for your reply and recommendations!
I have already used RobBERT. I just want to explore more models trained on the Dutch language. I will also try jina.