CMLM/ZhongJing-2-1_8b
Adapter vs Qwen/Qwen1.5-1.8B-Chat. CMLM/ZhongJing-2-1_8b is a LoRA adapter, not full weights: rank 8, on q_proj, v_proj in 24 layers (48 adapted matrices). It modifies Qwen/Qwen1.5-1.8B-Chat at load time.
Adapter vs Qwen/Qwen1.5-1.8B-Chat. CMLM/ZhongJing-2-1_8b is a LoRA adapter, not full weights: rank 8, on q_proj, v_proj in 24 layers (48 adapted matrices). It modifies Qwen/Qwen1.5-1.8B-Chat at load time.