netcat420/MFANN3bv0.20.18
Partial fine-tune vs liminerity/Phigments12. Partial change: 26% of the weights changed (attention, mlp, norms), in 32 of 32 layers.
Tensors changed: Norms 6/66, MLP 37/128, Attention 155/256, lm_head 0/2, Embeddings 0/1
1 derivatives checked
- mradermacher/MFANN3bv0.20.18-i1-GGUF: Quantization