katanemo/Arch-Router-1.5B.gguf
Full fine-tune vs Qwen/Qwen2.5-1.5B-Instruct. katanemo/Arch-Router-1.5B.gguf is a GGUF (Q6_K) of different weights. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Embeddings 1/1, Norms 43/57, MLP 84/84, Attention 195/196