3 years ago

Category: Docker, Jenkins, github, automation, machine-learning

If Machine Learning is about training model and finding best predictions then automation is its ultimate catalyst. In Machine Learning we write some code run it finds metrics like accuracy, change some hyperparameters, in case of NN add some more layers and when metrics are satisfactory model is ready to use for prediction. Thats how generally machine learning looks like.

2) Job2: This is to check the container required for model training and starting it.

Another thing to note is that if model is wicked then this Job has been intentionally failed to trigger the Job3 of model training (exit 1 return job failure).