/DL_develop_to_deploy_workshop

4-5 days workshop from development to deploy (Deep Learning, tensorlfow, keras, mlflow, docker, ...)

Primary LanguageJupyter Notebook

Workshop on Deep Learning (develop to deploy for starters)

  • Switch between branches in order to see all developed versions of code

The purpose of this online series is to develop an image classifier using TensorFlow and Keras on a funny dataset Mosquito-on-human-skin we don't talk MNIST here :D.

We will discuss why we need to log every single experiment and how to do it easily with MLflow with 2-3 lines of code, and then we will use the advanced functions of MLflow to expand our knowledge.

Here are some questions we'll try to answer:

  • MLflow: what is it? (Components, configuration, storage options, etc.)
  • Can you tell me why we need it? When and how do we use it?
  • Can you tell me what the benefits are?

The next step is to deploy the model using either MLflow Models or Tensorflow Serving, which are the two popular approaches.

A hands-on experience with basic optimization steps will be provided to help us understand why and when we need optimization.

It would be nice if we had enough time to develop an API (Application Programming Interface) using FastApi, and we would also need to use Docker so that maintenance on production become as simple as possible.

What's next? Could it be TFX?

A basic understanding of deep learning and machine learning is required