DavidAU/Qwen3.5-4B-Deckard-HERETIC-UNCENSORED-Thinking
LoRA-style vs coder3101/Qwen3.5-4B-heretic. 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, out_proj, q_proj, up_proj, v_proj changed, in 32 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/4, Norms 0/203, Attention 56/320, MLP 96/196
3 derivatives checked
- mradermacher/Qwen3.5-4B-Deckard-HERETIC-UNCENSORED-Thinking-i1-GGUF: Quantization
- mradermacher/Qwen3.5-4B-Deckard-HERETIC-UNCENSORED-Thinking-GGUF: Same weights, converted
- stardude1/Qwen3.5-4B-Deckard-HERETIC-UNCENSORED-Thinking-mlx-4Bit: Quantization