Minthy/RouWei-0.8
Full fine-tune vs Minthy/RouWei-0.7. Fine-tune: 97% of the weights changed, across 11 of 12 layers.
Tensors changed: other 534/662, Embeddings 571/590, Attention 700/716, Norms 442/546
1 derivatives checked
- John6666/mature-ritual-illustrious-v12-sdxl: Different architecture