cross-encoder/stsb-roberta-large

Full fine-tune vs FacebookAI/roberta-large. Fine-tune: 100% of the weights changed, across 24 of 24 layers.

Tensors changed: Norms 96/96, Attention 192/192, other 98/98, Embeddings 5/6

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse