- Ensure both Dockerfile and Dockerfile.gpu include latest main.py
- Picks up all syntax fixes for production deployment
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Switch to GPU-enabled Docker image
Andrew McCracken+1: Claude
2f6841c
- Using techdaskalos/cybersecchatbot:gpu from Docker Hub
- Enable GPU acceleration with N_GPU_LAYERS=35
- Configure MODEL_POOL_SIZE=10 for concurrent users
- Requires GPU hardware in HF Space settings
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Disable RAG for faster inference
Andrew McCracken+1: Claude
2a55dc3
- Set USE_RAG=false to remove knowledge base overhead
- Should significantly reduce response time on CPU instance
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Add test interface HTML to Docker image
Andrew McCracken+1: Claude
403caa4
- Copy test_interface.html to /app/ in Docker image
- Fixes 404 error on /test endpoint
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Add knowledge_db directory and remove deprecated env var
Andrew McCracken+1: Claude
7942e87
- Create /app/knowledge_db directory with write permissions
- Remove deprecated TRANSFORMERS_CACHE env var (use HF_HOME instead)
- Fixes PermissionError: [Errno 13] Permission denied: './knowledge_db'
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Fix Hugging Face cache permissions
Andrew McCracken+1: Claude
b721fc6
- Set HF_HOME, TRANSFORMERS_CACHE, HF_HUB_CACHE to /data/huggingface
- Create all required cache directories with proper permissions
- Fixes RuntimeError: Data processing error: I/O error: Permission denied
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Fix /app/models directory permissions
Andrew McCracken+1: Claude
3b792bc
- Create /app/models directory with write permissions
- Fixes PermissionError: [Errno 13] Permission denied: './models'
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Fix /data directory permissions for HF Spaces
Andrew McCracken+1: Claude
28a4990
- Create /data directory with write permissions
- Fixes sqlite3.OperationalError: unable to open database file
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Use pre-built Docker image from Docker Hub
Andrew McCracken+1: Claude
cf74856
- Updated Dockerfile to use techdaskalos/cybersecchatbot:latest
- Removed install_llama.py and start.sh (no longer needed)
- Updated requirements.txt to use llama_cpp_python without version pin
- Simplified deployment using pre-built AMD64 image
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Use pre-built Docker image from Docker Hub
Andrew McCracken
ccc0289
Fix: Install llama-cpp-python at startup to /tmp to avoid build timeout
Andrew McCracken
0a1ff0d
Fix: Install cmake and build-essential for llama-cpp-python build
Andrew McCracken
8e854ef
Fix: Force use of pre-built llama-cpp-python wheels
Andrew McCracken
a922ca8
Fix: Install llama-cpp-python at runtime (HF Spaces workaround)
Andrew McCracken
3fad655
Optimize: Use pre-built llama-cpp-python wheels for faster builds
