llmfan46/gemma-4-12B-it-qat-q4_0-uncensored-heretic-NVFP4-GGUF
Same weights, converted (fewer layers) vs llmfan46/gemma-4-12B-it-qat-q4_0-unquantized-uncensored-heretic. llmfan46/gemma-4-12B-it-qat-q4_0-unquantized-uncensored-heretic matches part of llmfan46/gemma-4-12B-it-qat-q4_0-uncensored-heretic-NVFP4-GGUF (618 of its 995 tensors; the rest of llmfan46/gemma-4-12B-it-qat-q4_0-uncensored-heretic-NVFP4-GGUF has no counterpart). On the shared part: Same weights stored in another dtype (BF16 → T40/BF16), plus 377 new tensors.
Tensors changed: Embeddings 0/1, Norms 0/289