Hacker News
points by lofties 3 days ago | 0 comments
BERT models are still amazing for what they do. Even the latest JEV/KEV/SEV/BEV/REV/TREV models can't beat a well-trained BERT.
[–] endymi0n 2 days ago
hehe, so much in a name... just today I told Claude off for recommending me a BERT model for voice identification and it told me it was actually SOTA 2025!

https://arxiv.org/html/2510.04213v1

[–] throwaw12 3 days ago
do you have examples of open weight BERT models from hugging face which are trained for specific use cases?
[–] rockinghigh 2 days ago
For these small models, you would fine tune them on your use case. https://huggingface.co/transformers/v4.8.2/training.html