Vortex5/Radiant-Shadow-12B
Partial fine-tune vs Retreatcost/KansenSakura-Radiance-RP-12b. Partial change: 65% of the weights changed (attention, mlp, norms), in 26 of 40 layers.
Tensors changed: Norms 50/81, MLP 78/120, Attention 104/160
2 derivatives checked
- mradermacher/Radiant-Shadow-12B-i1-GGUF: Quantization
- mradermacher/Radiant-Shadow-12B-GGUF: Quantization