boxwrench/gemma-4-qat-mtp-assistant-heads
Full fine-tune vs google/gemma-4-12B-it-qat-q4_0-unquantized-assistant. boxwrench/gemma-4-qat-mtp-assistant-heads is a GGUF (Q8_0) of different weights. Fine-tune: 100% of the weights changed, across 4 of 4 layers.
Tensors changed: Embeddings 1/1, Norms 4/13, MLP 12/12, Attention 8/8, other 2/2