Daizee/Gemma3-Callous-Calla-4B
LoRA-style vs google/gemma-3-4b-it. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 34 of 34 layers; norms, biases and embeddings are untouched.
Tensors changed: Norms 0/316, MLP 83/210, Attention 136/352, other 0/1, Embeddings 0/3
2 derivatives checked
- Daizee/Dirty-Calla-4B: LoRA-style
- ZeroXClem/Gemma3-4B-Arceus-Servant: LoRA-style