FIDLE Training


Since 2021, the CNRS institute in France is providing free trainings to acquire the basic skills and understanding of Deep Learning technology (1).

These courses are all freely accessible on YouTube and provide 2 to 3 hours of training on a variety of subjects related to AI technology.

This resource will be dedicated to explain a little deeper what are the training provided here for free.

History & Description of FIDLE

For many years, Dr. Jean-Luc Parouty, researcher at SIMaP lab in Grenoble, France (which is part of the French Institution CNRS), was in charge of internal training for CNRS researchers nationwide on AI technologies, from machine learning to deep learning algorithms.

After the COVID crisis hit, Dr. Parouty started to proposed his training in a video format in order to fulfill the trainings even during the different lock-downs.

Starting in 2021, the idea developed into the FIDLE project (Formation Introduction au Deep LEarning) with the help of Grenoble MIAI (Multidisciplinary Institute of Artificial Intelligence), the CNRS and the Grenoble University (UGA) as well as the IDRIS and the MITI departments of CNRS.

The FIDLE project is a based on YouTube videos, each about one AI technology that is explained during approximately 2 hours in a live stream every Thursday at 2pm (GMT+1). Each sequence is composed of a theorical part and practical course containing exercices on Jupyter lab notebooks (the notebooks are available offline and can be modified afterward, if needed be).

In order to be in the best conditions a FIDLE environment has been developed and can be installed (or dockerized) (2). a Wiki with detailed procedures is also available (3).

All the supports are in english but the presentations and the Wiki are in French only, for now.

The training are seasonal with a program evolving each year, depending on the subscribers requests and the AI technologies developed throughout the year. One season has approximately 20 episodes.

All episodes for every seasons are freely available on YouTube after the initial live stream and the support documents are also available for subscribers that still want to use them

The Second Season (2022/2023) is focused on:

  • Linear Regression in Neural Network
  • Convolutional Network
  • Recurrent Neural Network
  • PyTorch, TensorFlow
  • Transformers
  • Graph Neural Network
  • AutoEncoder and Variational AutoEncoder
  • Generative Adversarial Network
  • Diffusion Model
  • Ethical questions about AI
  • Deep Reinforcement Learning
  • Physics-Informed Neural Network

The First Season from 2021/2022 can be found here: