soob3123/amoral-gemma3-1B-v2
LoRA-style vs google/gemma-3-1b-it. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj, o_proj, q_proj, v_proj changed, in 26 of 26 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/157, MLP 0/78, Attention 104/104
4 derivatives checked
- mradermacher/amoral-gemma3-1B-v2-GGUF: Same weights, converted
- mradermacher/amoral-gemma3-1B-v2-i1-GGUF: Quantization
- soob3123/amoral-gemma3-1B-v2-gguf: Same weights, converted
- DK4AAD/amoral-gemma3-1B-v2-gguf: Same weights, converted