taozi555/NVIDIA-Nemotron-3-Nano-30B-A3B-novelrp-BF16
Full fine-tune vs unsloth/Nemotron-3-Nano-30B-A3B. Fine-tune: 94% of the weights changed, across 52 of 52 layers.
Tensors changed: other 23/207, Embeddings 0/1, Norms 0/76, MLP 2689/2691, Attention 24/24, lm_head 0/1