nineninesix/gepard-1.1
LoRA-style vs nineninesix/gepard-1.0. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of q_proj, v_proj changed, in 12 of 14 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/37, other 0/70, Norms 0/64, MLP 0/51, Attention 28/72