The purpose of this project is to test capabilities of the TensorFlow Serving. It should host a GAN model to predict Street View House Numbers.
Mentioned GAN model is trained using semi-supervised learning technique and bases on the model taught at Udacity Deep Learning Foundations course. The original Jupyter Notebook can be found at GitHub.
Additionally I made it possible to export trained model into Protobuf and implemented a client to request the prediction for a number image.
More details on the project can be found at my blog.
jackbbhua/tf_serving_example
Deployment of TensorFlow models into production with TensorFlow Serving, Docker, Kubernetes and Microsoft Azure
Python