yimn-Aghosh/zegrate-ai
LoRA-style vs Qwen/Qwen2.5-14B-Instruct. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of q_proj, v_proj changed, in 48 of 48 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/97, MLP 0/144, Attention 96/336, lm_head 0/1