nguyennm1024/dana-nemo-4b
LoRA-style vs nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 21 of 42 layers; norms, biases and embeddings are untouched.
Tensors changed: other 0/215, MLP 34/102, Embeddings 0/1, Norms 0/64, Attention 16/16, lm_head 0/1