Mitchins/Aronnax-N9Bv2-1.0
LoRA-style vs nvidia/NVIDIA-Nemotron-Nano-9B-v2. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of q_proj, v_proj changed, in 3 of 56 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, other 0/189, Norms 0/84, MLP 0/50, Attention 3/16, lm_head 0/1
2 derivatives checked
- mradermacher/Aronnax-N9Bv2-1.0-i1-GGUF: Quantization
- mradermacher/Aronnax-N9Bv2-1.0-GGUF: Same weights, converted