robertsw/aesthetics_v2
Full fine-tune vs facebook/dinov2-large. Fine-tune: 100% of the weights changed, across 24 of 24 layers.
Tensors changed: Embeddings 4/5, Attention 192/192, other 48/48, MLP 96/96, Norms 98/98
Full fine-tune vs facebook/dinov2-large. Fine-tune: 100% of the weights changed, across 24 of 24 layers.
Tensors changed: Embeddings 4/5, Attention 192/192, other 48/48, MLP 96/96, Norms 98/98