/Gather-Tensorflow-Serving

Gathers how to deploy tensorflow models using nginx, hadoop, kafka, flask, gunicorn, socketio, docker swarm, luigi spotify, airflow, celery and so much more!

Primary LanguagePythonMIT LicenseMIT

Gather-Tensorflow-Serving

Gather how to deploy tensorflow models as much I can

Covered

  1. Object Detection. Flask SocketIO + WebRTC
  2. Object Detection. Flask SocketIO + opencv
  3. Speech streaming. Flask SocketIO
  4. Text classification. Flask + Gunicorn
  5. Image classification. TF Serving
  6. Image Classification using Inception. Flask SocketIO
  7. Object Detection. Flask + opencv
  8. Face-detection using MTCNN. Flask SocketIO + opencv
  9. Face-detection using MTCNN. opencv
  10. Image classification using Inception. Flask + Docker
  11. Image classification using Inception. Flask + EC2 Docker Swarm + Nginx load balancer
  12. Text classification. Hadoop streaming MapReduce
  13. Text classification. Kafka
  14. Text classification. Distributed TF using Flask + Gunicorn + Eventlet
  15. Text classification. Tornado + Gunicorn
  16. Text classification. Flask + Celery + Hadoop
  17. Text classification. Luigi scheduler + Hadoop
  18. Text classification. Luigi scheduler + Distributed Celery
  19. Text classification. Airflow scheduler + elasticsearch + Flask
  20. Text classification. Apache Kafka + Apache Storm

Technology used

  1. Flask
  2. Flask SocketIO
  3. Gunicorn
  4. Eventlet
  5. Tornado
  6. Celery
  7. Hadoop
  8. Kafka
  9. Nginx
  10. WebRTC
  11. Luigi Spotify
  12. Airflow
  13. Elastic search
  14. Apache Storm

Printscreen

alt text

All folders contain print screens, logs and instructions on how to start.

Notes

  1. Deploy them on a server, change local in code snippets to your own IP.
  2. WebRTC chrome only can tested on HTTPS server.
  3. When come to real deployment, always prepare for up-scaling architectures. Learn about DevOps.
  4. Please aware with your cloud cost!