About TWESD | TWESD workshops  | Contact Us

The main objective of this training is to familiarize participants with the embedding concept of machine learning algorithms, in particular deep learning models on edge devices. This training includes software programming as well as Hardware implementation on an embedded platform. The training will be conducted over three sessions with hands on lab.

Objectives:

• Enable participants to train and convert machine learning models to TF lite format

• Educate participants on how to apply transfer learning techniques on TF lite models.

• Empower participants to develop and implement their own embedded AI projects

The program of the training is as follows:

• Session1 (8:30-10:30): Design and build models
1. Data preparation
2. Model building (TF/Keras)
3. Model training (custom dataset)
4. Model conversion (TF lite)
5. Comparative evaluation (Quantization/ Pruning/…etc)
• Session 2 (10:45-12:45): Transfer learning
1. Data preparation
2. Pre-trained model retrieval
3. Transfer learning
• Session3 (14:00-16:30): RaspberryPi implementation
1. Setting up a development environment on the RaspberryPi.
2. TF lite model implementation
3. Performance evaluation

For more details, you can download the Call for Training.
To participate, please fill in this use form.

 

---