nvidia/Hymba-1.5B-Instruct
Full fine-tune vs nvidia/Hymba-1.5B-Base. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 1/1, Norms 167/225, other 289/289, MLP 96/96
Full fine-tune vs nvidia/Hymba-1.5B-Base. Fine-tune: 100% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 1/1, Norms 167/225, other 289/289, MLP 96/96