dahara1/gemma-4-12B-it-qat-assistant_gguf

Same weights, converted vs google/gemma-4-12B-it-qat-q4_0-unquantized-assistant. Same weights stored in another dtype (BF16 → F16/F32), plus 5 new tensors.

Tensors changed: Embeddings 0/1, Norms 0/21, MLP 0/12, Attention 0/8, other 0/2

On the Hub · Downloads on Model Pulse