feat: add local LLM installation option with H100 GPU support

- Add Step 4.5: Local LLM Installation (Optional) to setup guide
- Implement H100 GPU resource checking (4x ideal, 1x minimum)
- Add NVIDIA API catalog navigation guide
- Implement automatic deployment strategy (local vs cloud fallback)
- Add Docker Compose integration for local NIM installation
- Auto-configure .env based on deployment choice
- Support multi-GPU configuration (1-4 H100 GPUs)
- Add verification and health checks for local installation

Author: T-DevH