Elfrino/SpaceGnome-V1
Full fine-tune vs Undi95/MXLewd-L2-20B. Fine-tune: 99% of the weights changed, across 62 of 62 layers.
Tensors changed: Embeddings 1/1, Norms 123/125, MLP 184/186, Attention 244/248, lm_head 1/1
1 derivatives checked
- mradermacher/SpaceGnome-V1-i1-GGUF: Quantization