-
refactor: clean project structure to essentials only
Marcos +1: Claude
- Renamed main server to server.py (single server)
- Created simplified start.sh script
- Removed 200+ redundant files:
- 60+ duplicate test files
- Unused directories (llama-omni2-official, llama_omni2_integration)
- Old installation scripts
- Development/debug files
- Kept only essential files:
- server.py (main WebRTC server with GPU+vLLM)
- install.sh (complete installation with exact versions)
- start.sh (simple startup script)
- llama_omni2/ (core implementation)
- requirements.txt
- Model: Qwen3-0.6B with vLLM 0.8.4
- Performance: 378ms latency, 92% coherence
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
