DEPLOYMENT.md

Diamond CSGO AI Player - Deployment Guide

🚀 Deploying to Hugging Face Spaces

Prerequisites

1. Hugging Face account
2. Model checkpoint files (agent_epoch_00206.pt or similar)
3. Git and Git LFS installed

Step 1: Prepare Repository

1. Clone/Fork this repository

2. Install Git LFS (for large model files):

   git lfs install
   git lfs track "*.pt"
   git add .gitattributes
   

3. Add your model checkpoint:

   # Copy your trained model to the project root
   cp /path/to/your/agent_epoch_00206.pt .
   git add agent_epoch_00206.pt
   git commit -m "Add trained model checkpoint"
   

Step 2: Create Hugging Face Space

1. Go to Hugging Face Spaces
2. Click "Create new Space"
3. Configure:
   • Space name: diamond-csgo-ai (or your choice)
   • License: Your preferred license
   • Space SDK: Docker
   • Space hardware:
     • CPU basic (free) - for demo/testing
     • GPU T4 small (paid) - for better performance

Step 3: Upload Code

# Clone your new space
git clone https://huggingface.co/spaces/YOUR_USERNAME/YOUR_SPACE_NAME
cd YOUR_SPACE_NAME

# Copy all project files
cp -r /path/to/diamond/* .

# Commit and push
git add .
git commit -m "Initial Diamond CSGO AI deployment"
git push

Step 4: Configuration Files

Ensure these files are in your space root:

• app.py - Main FastAPI application
• requirements.txt - Python dependencies
• Dockerfile - Container configuration
• README.md - Space description
• packages.txt - System packages (if needed)

Step 5: Model Setup

If your model is too large for Git (>100MB), use Git LFS or download from Hub:

# In your app.py, add model downloading:
from huggingface_hub import hf_hub_download

def download_model():
    return hf_hub_download(
        repo_id="YOUR_USERNAME/YOUR_MODEL_REPO",
        filename="agent_epoch_00206.pt"
    )

🔧 Local Testing

Before deploying, test locally:

# Install dependencies
pip install -r requirements.txt

# Run tests
python test_web_app.py

# Start local server
python run_web_demo.py

Visit http://localhost:7860 to test the interface.

⚙️ Configuration Options

Hardware Requirements

Tier	CPU	RAM	GPU	Performance
Free	2 vCPU	16GB	None	Basic demo
Basic GPU	4 vCPU	16GB	T4	Good performance
Premium	8 vCPU	32GB	A10G	Best experience

Environment Variables

Add these in your Space settings:

• CUDA_VISIBLE_DEVICES="" (for CPU-only)
• PYTHONPATH="/app/src:/app"

🎮 Usage Instructions

Once deployed, users can:

1. Visit your Space URL
2. Click on the game canvas
3. Use keyboard controls:
   • WASD - Movement
   • Space - Jump
   • Arrow keys - Camera
   • 1,2,3 - Weapons
   • R - Reload
   • M - Switch Human/AI mode

🐛 Troubleshooting

Common Issues

1. Model not loading:

   • Check checkpoint file exists
   • Verify file size (<5GB for Spaces)
   • Use Git LFS for large files

2. Import errors:

   • Check requirements.txt is complete
   • Verify Python path in Dockerfile

3. Performance issues:

   • Use GPU hardware tier
   • Reduce model complexity
   • Lower frame rate

4. WebSocket connection failed:

   • Check firewall settings
   • Verify port 7860 is accessible
   • Try different browser

Debug Mode

Enable debug logging:

import logging
logging.basicConfig(level=logging.DEBUG)

📊 Monitoring

Monitor your Space:

• View logs in HF Spaces interface
• Check GPU utilization
• Monitor user sessions

🔄 Updates

To update your deployed Space:

git pull  # Get latest changes
git add .
git commit -m "Update to latest version"
git push  # Automatically redeploys

💡 Tips for Success

1. Start with CPU tier to test basic functionality
2. Use smaller models for faster loading
3. Test thoroughly locally before deploying
4. Monitor resource usage to optimize costs
5. Add usage instructions in your Space README

🎯 Next Steps

After successful deployment:

• Share your Space with the community
• Collect user feedback
• Iterate on the interface
• Add new features like replay saving
• Consider multi-user support

Happy deploying! 🚀