simpledirect/Vinci-MLE-30B-1.0
LoRA-style vs ibm-granite/granite-4.1-30b. 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 64 of 64 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/129, MLP 190/192, Attention 250/256
1 derivatives checked
- simpledirect/Vinci-MLE-30B-1.0-GGUF: Same weights, converted