ermiaazarkhalili/Qwen3.8-4B-SFT-Fable5
LoRA-style vs empero-ai/Qwen3.8-4B-Distill. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, in_proj_a, in_proj_b, in_proj_qkv, in_proj_z, 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 0/5, Norms 0/209, Attention 152/324, MLP 96/199, other 0/1
1 derivatives checked
- mradermacher/Qwen3.8-4B-SFT-Fable5-GGUF: Same weights, converted