Exercises for the Deep Learning course held at FEI, VSB-TU Ostrava.

Course information may be found here.

You can find more details about the course at my homel.

Feel free to contact me if you have any questions or want to discuss any topic from the course 😊

All authorship is mentioned where possible.

📊 Exercises

Exercise 1

The aim of the exercise is to get an overview of the basic capabilities of the Keras library and to build a simple neural network for MNIST dataset classification.

Jupyter Notebook

Google Colab

Exercise 2

The goal of the exercise is to learn how to solve regression problems using deep learning.

Jupyter Notebook

Google Colab

Exercise 3

The aim of the exercise is to learn how to use the basic architecture based on convolutional layers and how to classify image data.

Jupyter Notebook

Google Colab

Exercise 4

The aim of the exercise is to learn how to use transfer learning for image data, in the second part of the exercise we will look at time series classification using CNN.

Jupyter Notebook

Google Colab

Exercise 5

The goal of the exercise is to learn how to use Autoencoder and Variational autoencoder architectures in image data to generate new image data instances and detect anomalies.

Jupyter Notebook

Google Colab

Exercise 6

The goal of this exercise is to learn how to use recurrent neural networks for sentiment analysis of text data. In the exercise, we will work with data from Twitter.

Jupyter Notebook

Google Colab

Exercise 7

The goal of this exercise is to learn how to create your own Word2Vec embedding and generate your own text using RNN.

Jupyter Notebook

Google Colab

Exercise 8

The aim of the exercise is to learn how to work with the Attention mechanism within the RNN.

Jupyter Notebook

Google Colab

Exercise 9

The exercise focuses on the use of transformer models for text data classification using the Hugging face API.

Jupyter Notebook

Google Colab

Exercise 10

This lecture is focused on using CNN for object localization tasks, we will also use YOLOv8 model in a real-world scenario during second part of the lecture.

Jupyter Notebook

Google Colab

Exercise 11

The aim of the lecture is to get an overview of possibilities in the generative artificial intelligence (GenAI) domain

Jupyter Notebook - Stable Diffusion

Google Colab - Stable Diffusion

Jupyter Notebook - LLM

Google Colab - LLM

💡 Notes

How to use Kaggle notebooks

• You can use (Kaggle)[https://www.kaggle.com/] as an alternative to Google Colab
  • 📌 Beware that both platforms use different configuration and libraries versions thus full compatibility cannot be always guaranteed
• For importing the Jupyter notebook perform these steps:
  • Click on + sign (or Create) button in the left panel and select New Notebook
  • In the new notebook select File > Import notebook > Link and paste URL of the Jupyter notebook from Github
    • e.g. https://github.com/rasvob/VSB-FEI-Deep-Learning-Exercises/blob/main/dl_01.ipynb
  • In the Notebook sidebar (right side, it can be expanded through small arrow icon in the bottom right corner) use these Session options:
    • Accelerator: GPU T4x2 or GPU P100
    • Persistence: Variables and Files
  • Own datasets can be uploaded using the the Notebook sidebar as well - Input section
    • Click on Upload > New dataset > File and Drag&Drop your file(s)
    • Set the Dataset title and click on Create
      • 💡 zip archives are automatically extracted
      • You can copy path of the file using the copy icon when you hover over the filename
        • The usual path is in format /kaggle/input/<dataset_name>/<filename>
  • 💡 There is some problem with using the hdf5 format in the filepath parameter in ModelCheckpoint
    • Use filename best.weights.h5 instead (hdf5 and h5 is the same format)
    • 💡 Remember to change the path in the load_weights() function as well!**
  • You can download your .ipynb notebooks using File > Download notebook option

How to create a Python Virtual Enviroment named venv

Create venv

python -m venv venv

Activate venv

• Activate venv in Windows

.\venv\Scripts\Activate.ps1

• Activate venv in Linux

source venv/bin/activate

Intall python packages

• Works for tensorflow 2.15.0

pip install jupyter "jupyterlab>=3" "ipywidgets>=7.6"
pip install pandas matplotlib requests seaborn scipy scikit-learn optuna scikit-image pyarrow opencv-python plotly==5.18.0 tensorflow[and-cuda] nltk textblob transformers datasets huggingface_hub evaluate

Test TF2 installation

• It should print list of all your GPUs
  • 💡 It is not working if an empty list [] is printed

python3 -c "import tensorflow as tf; print(tf.config.list_physical_devices('GPU'))"