dleemiller/finecat-nli-l

Full fine-tune vs answerdotai/ModernBERT-large. Fine-tune: 100% of the weights changed, across 28 of 28 layers.

Tensors changed: other 1/1, Norms 57/57, Embeddings 2/2, Attention 56/56, MLP 56/56

1 derivatives checked

• dleemiller/crossingguard-nli-l: Full fine-tune

On the Hub · Downloads on Model Pulse