UmbrellaInc/Wesker-Project-3.2-1B
LoRA-style vs D1rtyB1rd/Dirty-Alice-RP-NSFW-llama-3.2-1B. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 16 of 16 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/1, Norms 0/33, MLP 48/48, Attention 62/64
2 derivatives checked
- mradermacher/Wesker-Project-3.2-1B-i1-GGUF: Quantization
- mradermacher/Wesker-Project-3.2-1B-GGUF: Same weights, converted