A propos

This repository contains all the documents and links of the Fidle Training.

The objectives of this training, co-organized by the Formation Permanente CNRS and the SARI and DEVLOG networks, are :

• Understanding the bases of deep learning neural networks (Deep Learning)
• Develop a first experience through simple and representative examples
• Understand the different types of networks, their architectures and their use cases.
• Understanding Tensorflow/Keras and Jupyter lab technologies on the GPU
• Apprehend the academic computing environments Tier-2 (meso) and/or Tier-1 (national)

Course materials

Get the course slides

Useful information is also available in the wiki

Jupyter notebooks :

1. Linear regression with direct resolution
         Direct determination of linear regression
2. Linear regression with gradient descent
         An example of gradient descent in the simple case of a linear regression.
3. Complexity Syndrome
         Illustration of the problem of complexity with the polynomial regression
4. Logistic regression, in pure Tensorflow
         Logistic Regression with Mini-Batch Gradient Descent using pure TensorFlow.
5. Regression with a Dense Network (DNN)
         A Simple regression with a Dense Neural Network (DNN) - BHPD dataset
6. Regression with a Dense Network (DNN) - Advanced code
         More advanced example of DNN network code - BHPD dataset
7. CNN with GTSRB dataset - Data analysis and preparation
         Episode 1: Data analysis and creation of a usable dataset
8. CNN with GTSRB dataset - First convolutions
         Episode 2 : First convolutions and first results
9. CNN with GTSRB dataset - Monitoring
         Episode 3: Monitoring and analysing training, managing checkpoints
10. CNN with GTSRB dataset - Data augmentation
          Episode 4: Improving the results with data augmentation
11. CNN with GTSRB dataset - Full convolutions
          Episode 5: A lot of models, a lot of datasets and a lot of results.
12. CNN with GTSRB dataset - Full convolutions as a batch
          Episode 6 : Run Full convolution notebook as a batch
13. Tensorboard with/from Jupyter
          4 ways to use Tensorboard from the Jupyter environment
14. Text embedding with IMDB
          A very classical example of word embedding for text classification (sentiment analysis)
15. Text embedding with IMDB - Reloaded
          Example of reusing a previously saved model
16. Text embedding/LSTM model with IMDB
          Still the same problem, but with a network combining embedding and LSTM

Installation

To run this examples, you need an environment with the following packages :

• Python >3.5
• numpy
• Tensorflow 2.0
• scikit-image
• scikit-learn
• Matplotlib
• seaborn
• pyplot

You can install such a predefined environment :

conda env create -f environment.yml

To manage conda environment see there

Licence

[en] Attribution - NonCommercial - ShareAlike 4.0 International (CC BY-NC-SA 4.0)
[Fr] Attribution - Pas d’Utilisation Commerciale - Partage dans les Mêmes Conditions 4.0 International See License.
See Disclaimer.