- Add --no-ram CLI option to force model loading without CPU RAM spilling
- Implement --no-ram behavior for:
  - llama-cpp-python: n_gpu_layers=-1, use_mmap=False, ignore --n-ctx
  - HuggingFace transformers: device_map='cuda:0', low_cpu_mem_usage=True
  - Diffusers: force full GPU loading
  - sd.cpp: maximize GPU usage
- Propagate flag through model manager
- Add startup banner message