realrebelai/KREA-2_GGUFs
Full fine-tune vs krea/Krea-2-Turbo. realrebelai/KREA-2_GGUFs is a GGUF (Q8_0) of different weights. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Norms 129/129, other 35/35, MLP 105/105, Attention 160/160