DireDreadlord/GemCod-Zircon-270M-XM
LoRA-style vs google/gemma-3-270m-it. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj, o_proj, q_proj, v_proj changed, in 18 of 18 layers; norms, biases and embeddings are untouched.
Tensors changed: Norms 0/109, MLP 0/54, Attention 72/72
1 derivatives checked
- mradermacher/GemCod-Zircon-270M-XM-GGUF: Same weights, converted