almanach/ModernBERT-bio-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
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