DahonoLabs/Dahono-4B
LoRA-style vs Qwen/Qwen3.5-4B-Base. 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 31 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/5, MLP 96/199, Attention 32/324, Norms 0/209, other 0/1
1 derivatives checked
- mradermacher/Dahono-4B-GGUF: Same weights, converted
1 other repos hold byte-for-byte identical weight files.