s-nlp/bart-base-detox
Full fine-tune vs facebook/bart-base. Fine-tune: 98% of the weights changed, across 6 of 6 layers.
Tensors changed: Embeddings 4/6, Attention 144/144, Norms 60/60, MLP 48/48
Full fine-tune vs facebook/bart-base. Fine-tune: 98% of the weights changed, across 6 of 6 layers.
Tensors changed: Embeddings 4/6, Attention 144/144, Norms 60/60, MLP 48/48