thomas-sounack/BioClinical-ModernBERT-large
Full fine-tune vs answerdotai/ModernBERT-large. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: other 2/2, Norms 57/57, Embeddings 2/2, Attention 56/56, MLP 56/56
4 derivatives checked
- OpenMed/OpenMed-PII-Spanish-BioClinicalModern-Large-395M-v1: Full fine-tune
- OpenMed/OpenMed-PII-French-BioClinicalModern-Large-395M-v1: Full fine-tune
- OpenMed/OpenMed-PII-Italian-BioClinicalModern-Large-395M-v1: Full fine-tune
- OpenMed/OpenMed-PII-German-BioClinicalModern-Large-395M-v1: Full fine-tune