netcat420/MFANN3b-SFT-Rebase
LoRA-style vs liminerity/Phigments12. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj, q_proj, v_proj changed, in 31 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: Norms 0/66, MLP 0/128, Attention 93/256, lm_head 0/2, Embeddings 0/1
1 derivatives checked
- mradermacher/MFANN3b-SFT-Rebase-i1-GGUF: Quantization