kueizen/Marco-Mini-Instruct-REAP30
LoRA-style vs ATH-MaaS/Marco-Mini-Instruct. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, up_proj changed, in 28 of 28 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, MLP 1920/1939, Norms 0/113, Attention 0/112