Ryex/Floppa-12B-Gemma3-Uncensored
LoRA-style vs google/gemma-3-12b-it. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, fc1, fc2, gate_proj, k_proj, o_proj, out_proj, q_proj, up_proj, v_proj changed, in 48 of 48 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/4, Norms 0/400, MLP 198/252, Attention 300/408, other 0/1
4 derivatives checked
- mradermacher/Floppa-12B-Gemma3-Uncensored-i1-GGUF: Quantization (fewer layers)
- mradermacher/Floppa-12B-Gemma3-Uncensored-GGUF: Quantization (fewer layers)
- Ryex/Floppa-12B-Gemma3-Uncensored-Q5_K_M-GGUF: Quantization (fewer layers)
- QrCode99/Floppa-12B-Gemma3-Uncensored-Q8_0-GGUF: Quantization (fewer layers)
1 other repos hold byte-for-byte identical weight files.