Contribute some benchmark

#13
by TamThai - opened

I'm glad that you opened this awesome model ✨. The benchmark dataset contains about 100 samples related to automotive documents (AUTOSAR, ECUs), and gte-modernbert-base performed excellently. Manual validation gave me goosebumps because its retrieval always hit the top 3.

image.png

Alibaba-NLP org

Thank you so much for your attention to the GTE series of models. I'm happy to hear that the GTE-ModernBERT model has achieved very competitive performance on your benchmark.

Sign up or log in to comment