Deploying Your Tensorflow Model

Learn to deliver your models to millions using tensorflow serving by Rufai Mustapha.

Transcript

2. TensorFlow Playground! Deploying Your Model
3. Development Environment TensorFlow Serving Tensor Flow Game Development 01 02 03 04 Table of contents
4. Rufai! Mustapha Software Engineer
5. Tensorflow
6. What is TensorFlow? A) A high-level programming language B) A cloud computing platform C) An open-source machine learning framework D) A database management system
7. Which of the following is NOT a core component of TensorFlow? A) Tensors B) Graphs C) Sessions D) Loops
8. What is the primary data structure in TensorFlow? A) Lists B) Dictionaries C) Tensors D) Arrays
9. Agenda Environment For Deployment Platforms Deployment Frameworks
10. Platform 00
11. Web
12. Web - First
13. Web - Iris
14. Android
15. Deployment Environment 01
18. apt-get remove tensorflow-model-server
19. echo "deb http://storage.googleapis.com/tensorflow-serving-apt stable tensorflow-model-server tensorflow-model-server-universal" | tee /etc/apt/sources.list.d/tensorflow-serving.list && \ curl https://storage.googleapis.com/tensorflow-serving-apt/tensorflow- serving.release.pub.gpg | apt-key add -
21. Deployment Architecture 02
22. A Web Deployment Architecture
23. A production requires more than code
24. A production requires more than code
25. TensorFlow Serving
26. What Is Serving?
27. TensorFlow Serving Flexibility
28. TensorFlow Serving Flexibility Compatibility
29. TensorFlow Serving Flexibility Compatibility High Performance
30. TensorFlow Serving Flexibility Compatibility High Performance Inference Framework
31. TensorFlow Serving Flexibility Compatibility High Performance Inference Framework GPU Acceleration
32. TensorFlow Serving Flexibility Compatibility High Performance Inference Framework Model Version Control and Monitoring
33. TensorFlow Serving Architecture
34. What Is Serving?
35. What is the main purpose of TensorFlow Serving? a) To train machine learning models b) To deploy and serve machine learning models in production c) To create neural networks d) To preprocess data for machine learning models
36. Which of the following is NOT a benefit of using TensorFlow Serving? a) Low latency and high throughput b) Supports multiple models and versions c) Automatically tunes hyperparameters of models d) Seamless integration with Docker and Kubernetes
37. TensorFlow Serving supports which two main communication protocols? a) HTTP and WebSockets b) RESTful API and gRPC c) SOAP and FTP d) TCP/IP and SSH
38. In TensorFlow Serving, which format is used to save a trained model for serving? a) CSV b) TensorFlow Lite c) SavedModel d) HDF5
39. Which of the following statements is TRUE about TensorFlow Serving's architecture? a) TensorFlow Serving handles model training and hyperparameter tuning b) TensorFlow Serving manages model loading, prediction requests, and model lifecycle c) TensorFlow Serving is only used for training models d) TensorFlow Serving requires models to be in the TensorFlow Lite format
41. References
42. Does anyone have any questions? Twitter @iafur https://mrufai.com