lucifer-ms/task-3-Qwen-Qwen1.5-1.8B
Adapter vs Qwen/Qwen1.5-1.8B. lucifer-ms/task-3-Qwen-Qwen1.5-1.8B is a LoRA adapter, not full weights: rank 4, on q_proj, v_proj in 24 layers (48 adapted matrices). It modifies Qwen/Qwen1.5-1.8B at load time.
Adapter vs Qwen/Qwen1.5-1.8B. lucifer-ms/task-3-Qwen-Qwen1.5-1.8B is a LoRA adapter, not full weights: rank 4, on q_proj, v_proj in 24 layers (48 adapted matrices). It modifies Qwen/Qwen1.5-1.8B at load time.