For people who struggle to start in deep learning with TensorFlow
This hands-on in-person workshop is based on Deep Learning with TensorFlow Course by IBM Cognitive Class
Learn how to get started with TensorFlow to capture relevant structure in images, sound, and textual data from unlabeled and unstructured data.
The workshop will cover core topics:
| Data Graph | Tensors | ReLu |
|---|---|---|
 |
 |
 |
- HelloWorld with TensorFlow
- Linear and Logistic Regression with TensorFlow
- Activation Functions
- Introduction to Convolutional Networks
- Convolution and Feature Learning
- Convolution with Python and TensorFlow
- MNIST Dataset
- Multilayer Perceptron with TensorFlow
- Convolutional Network with TensorFlow
| Sequentaial Data | Recurrent Model | LSTM |
|---|---|---|
 |
 |
 |
- Recurrent Neural Network Model
- Long Short-Term Memory
- Recursive Neural Tensor Network Theory
- Applying Recurrent Networks to Language Modelling
| Forward Pass | Backward Pass | Quality Assessment |
|---|---|---|
 |
 |
 |
- Applications of Unsupervised Learning
- Restricted Boltzmann Machine
- Training a Restricted Boltzman Machine
- Recommendation System with a Restrictive Boltzman Machine
| Encode/Decode | Architecture | Autoencoder vs RBM |
|---|---|---|
 |
 |
 |
- Introduction to Autoencoders and Applications
- Autoencoder Structure
- Deep Belief Network
- Python for Data Science Workshop
- Data Analysis with Python Workshop
- Machine Learning with Python Workshop
You will need a laptop that can access the internet
Install miniconda or install the (larger) Anaconda distribution
Install Python using Miniconda
OR Install Python using Ananconda
Clone the repository
git clone git@github.com:aymanibrahim/dltf.git
OR Download the repository as a .zip file
Change current directory to dltf directory
cd dltf
Install Python with the required packages into an environment named dltf as per environment.yml YAML file.
conda env create -f environment.yml
When conda asks if you want to proceed, type "y" and press Enter.
Change the current default environment (base) into dltf environment.
conda activate dltf
Install ipywidgets JupyterLab extension
jupyter labextension install @jupyter-widgets/jupyterlab-manager
Enable widgetsnbextension
jupyter nbextension enable --py widgetsnbextension --sys-prefix
Use check_environment.py script to make sure everything was installed correctly, open a terminal, and change its directory (cd) so that your working directory is the workshop directory dltf you cloned or downloaded. Then enter the following:
python check_environment.py
If everything is OK, you will get the following message:
Your workshop environment is set up
Start JupyterLab using:
jupyter lab
JupyterLab will open automatically in your browser.
You may access JupyterLab by entering the notebook server’s URL into the browser.
Press CTRL + C in the terminal to stop JupyterLab.
Change the current environment (dltf) into the previous environment.
conda deactivate
- Python: Programming language
- Conda: Package and environment manager
- Anaconda: Python distribution
- Miniconda: Minimal installer for conda
- NumPy: Fundamental package for scientific computing with Python
- Matplotlib: Python 2D plotting library
- seaborn: Statistical Data Visualization
- pandas: Python data analysis library
- scikit-learn: Machine Learning in Python
- TensorFlow: Deep Learning in Python
- Jupyter Notebook: Web application to create documents with code, equations, visualizations and text
- JupyterLab: Web-based development environment for Jupyter Notebooks
- Python for Data Science: Course by IBM Cognitive Class
- Data Analysis with Python: Course by IBM Cognitive Class
- Data Visualization with Python: Course by IBM Cognitive Class
- Machine Learning with Python: Course by IBM Cognitive Class
- Deep Learning with TensorFlow Course by IBM Cognitive Class
Thanks for your interest in contributing! There are many ways to contribute to this project. Get started here.
Deep Learning with TensorFlow Workshop by Ayman Ibrahim is licensed under a Creative Commons Attribution 4.0 International License. Based on a work at IBM Cognitive Class Deep Learning with TensorFlow Course by Saeed Aghabozorgi, PhD , Rafael Belo da Silva, Erich Natsubori Sato and Walter Gomes de Amorim Junior.