ibnsina-llm/ibnsina-30b-chat

Full fine-tune vs ibnsina-llm/ibnsina-30b. Fine-tune: 94% of the weights changed, across 48 of 48 layers.

Tensors changed: Embeddings 0/1, Norms 171/193, MLP 2609/2613, Attention 192/192, lm_head 1/1

2 derivatives checked