toiletsandpaper/russian-ai-text-detector-bert
Full fine-tune vs cointegrated/rubert-tiny2. Fine-tune: 100% of the weights changed, across 3 of 3 layers.
Tensors changed: Embeddings 5/5, Norms 12/12, Attention 24/24, other 14/14
Full fine-tune vs cointegrated/rubert-tiny2. Fine-tune: 100% of the weights changed, across 3 of 3 layers.
Tensors changed: Embeddings 5/5, Norms 12/12, Attention 24/24, other 14/14