mlx-community/gemma-4-12B-bf16

Same weights, repacked vs google/gemma-4-12B. Same weights in different files (re-saved, re-sharded or new metadata), plus 11 new tensors.

Tensors changed: Embeddings 0/1, Norms 0/289, other 0/48, MLP 0/144, Attention 0/184

On the Hub · Downloads on Model Pulse