Peter-A-P/smallprint-4b
LoRA-style vs google/gemma-4-E4B. 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 42 of 42 layers; norms, biases and embeddings are untouched.
Tensors changed: MLP 126/606, Norms 0/484, other 0/261, Attention 132/719, Embeddings 0/6
1 derivatives checked
- mradermacher/smallprint-4b-GGUF: Same weights, converted (fewer layers)