Retreatcost/Limn-Alpha-12B
LoRA-style vs Gryphe/Gemma-4-12B-StyleTune. 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 48 of 48 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/291, other 0/54, MLP 144/144, Attention 152/184
2 derivatives checked
- mradermacher/Limn-Alpha-12B-i1-GGUF: Quantization
- mradermacher/Limn-Alpha-12B-GGUF: Quantization