Omega-Factory/OLMoE-0125-Instruct-Compressed-Experimental
LoRA-style vs allenai/OLMoE-1B-7B-0125-Instruct. 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 16 of 16 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/65, MLP 2448/2448, Attention 64/64, lm_head 0/1