usermma/Mythos-nano-OBLITERATED
Full fine-tune vs squ11z1/Mythos-nano. Fine-tune: 95% of the weights changed, across 34 of 36 layers.
Tensors changed: Embeddings 1/1, Norms 0/73, MLP 102/108, Attention 136/252
2 derivatives checked
- mradermacher/Mythos-nano-OBLITERATED-i1-GGUF: Quantization
- mradermacher/Mythos-nano-OBLITERATED-GGUF: Same weights, converted