PRE_DEPLOYMENT_CHECKLIST.md

Pre-Deployment Checklist ✅

Complete this checklist before deploying to Hugging Face Spaces.

1. Model Files

• pytorch_model.pth is in backend/api/models/
• model_best.keras is in backend/api/models/
• Both models are in HuggingFace LFS format (.gitattributes)
• Models are not corrupted (test locally first)

2. Local Testing

• Cloned repository successfully
• Added trained models
• Built Docker image: docker build -t intel .
• Container starts without errors
• Can access http://localhost:7860
• Classification endpoint works: POST to /api/classify/
• Frontend loads properly
• API documentation visible at /swagger/

3. Configuration

• Updated .env if using custom settings
• Verified ALLOWED_HOSTS includes HF domains
• Checked CORS_ALLOWED_ORIGINS is correct
• Confirmed DEBUG=False for production
• Reviewed DJANGO_SECRET_KEY (auto-set is fine)

4. Code Quality

• No hardcoded API keys or passwords
• Removed development-only code
• All imports are available in requirements.txt
• No local file paths hardcoded
• Verified .gitignore doesn't exclude needed files

5. Git & Repository

• All changes committed: git status is clean
• .env is in .gitignore (no secrets in repo)
• Model files tracked with Git LFS if large
• Remote configured: git remote add hf <HF_SPACE_URL>
• Tested: git push --dry-run hf main

6. Documentation

• README.md updated with your username
• Updated citation with your info
• Added your model training notes if relevant
• Confirmed DEPLOYMENT.md is accurate
• Link to your HF Space is correct

7. Hugging Face Setup

• Created Hugging Face Space
• Set space as "Docker" SDK
• Set visibility (Public/Private)
• Added appropriate description and license
• Configured any secrets needed (if API keys used)

8. Final Tests (Optional)

• Run docker build . one more time
• Test with different image formats (jpg, png, webp)
• Test with both model options (pytorch, tensorflow)
• Verify error handling (invalid image, bad requests)
• Check performance with larger images

9. Deployment

• All above checks passed ✅
• Ready message: git push hf main
• Monitoring HF Space for build progress
• Verified Space URL is accessible
• Tested deployed Space (give it 5 minutes to start)

10. Post-Deployment

• Space is running without errors
• All endpoints respond correctly
• Frontend loads on live URL
• API returns correct predictions
• No CORS or security errors
• Shared link with team/public

---

Deployment Command

# Final push to Hugging Face
git push hf main

# View logs (from HF Space interface)
# Monitor at: https://huggingface.co/spaces/USERNAME/Intel_classification/logs

---

Rollback Plan

If something goes wrong:

# Revert to previous working state
git log --oneline | head -5
git reset --hard <COMMIT_HASH>
git push hf main --force  # Force push (use with caution)

# Or fix issues and push again
git add .
git commit -m "Fix: description of fix"
git push hf main

---

Common Issues & Fixes

Issue	Solution
Build timeout	Reduce image size or optimize dependencies
Out of memory	Prune unused Docker images/volumes
Models not loading	Verify file names match exactly
CORS errors	Check Django CSRF_TRUSTED_ORIGINS
Port conflicts	Ensure 7860 is available locally
Frontend not showing	Check React build in frontend/build/

---

Success Indicators ✨

Your deployment is successful when:

1. ✅ HF Space shows "Running" status
2. ✅ Website loads at your HF Space URL
3. ✅ Cannot refund with different images
4. ✅ Both models are selectable
5. ✅ API documentation (/swagger) is accessible
6. ✅ Admin panel works (/admin/)
7. ✅ No errors in HF Space logs

---

Need Help?

1. Documentation: DEPLOYMENT.md
2. Quick Start: QUICK_START.md
3. Changes Made: REORGANIZATION.md
4. GitHub Issues: Report bugs
5. HF Discussions: Ask community

---

Estimated Total Time: 15-30 minutes (first-time setup)

Good luck! 🚀