Machine Learning In Application

Practical implemantation of Machine Learning in the Application.

Architecture

Model: Machine Learning Model
Trainer: Training the model. So training process (loss, optimizer) is separated from Model.
Model API: Interface between the Model and Application.
DataProcessor: Load the data and preprocess it. It is used in Trainer and ModelAPI.
Resource: It manages the parameters for Trainer, Model and DataProcessor.

handwritten number recognizer by Chainer.

You can deploy this application by docker.

Please refer this tutorial to deploy the application.