iapp/openthai2.0-legal-thaillm-nemotron-3-nano-30b-a3b
Full fine-tune vs nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16. Fine-tune: 100% of the weights changed, across 52 of 52 layers.
Tensors changed: Embeddings 1/1, other 132/207, Norms 42/76, MLP 1584/1585, Attention 24/24, lm_head 1/1