thanhkt/Qwen2.5-3B-MultiChat-Vi
Adapter vs unsloth/Qwen2.5-3B-Instruct-bnb-4bit. thanhkt/Qwen2.5-3B-MultiChat-Vi is a LoRA adapter, not full weights: rank 16, on down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj in 36 layers (252 adapted matrices). It modifies unsloth/Qwen2.5-3B-Instruct-bnb-4bit at load time.