bugrabilge/Omni-31B-Turkish-Reasoning-Model

Full fine-tune vs google/gemma-4-31B. Fine-tune: 98% of the weights changed, across 60 of 60 layers.

Tensors changed: Embeddings 1/4, Norms 280/523, other 0/62, MLP 180/261, Attention 230/338

2 derivatives checked