Intro to Deep Learning

Jordan DeKraker & Haider Al-Tahan

This is the practical component of this tutorial on deep learning presented in Brain Hack Western 2019. Topics include:

1. Introduction to TensorFlow 2.0 and Keras.
2. Train and evaluate of dense and convolutional neural networks for real problems.
3. An open session for exploring and adapting existing NNs or for helping attendees build their own networks.

You can use Google Colaboratory notebook file which is a Jupyter notebook with free limited access to a GPU. Jupyter notebooks allows python programs to run directly in the browser which can be great way to learn and use TensorFlow.

Instructions

All material of the practical presentation is available in the jupyter notebook (<FileName>.ipynb). You can run jupyter notebook locally or using google colab.

• To run it on google colab,
  1. go to http://colab.research.google.com and click on File located in top-left corner > Open notebook....
  2. Click on Github in the header of the opened view.
  3. Copy and Paste this link haideraltahan/DeepLearningIntro-BrainHack.
  4. Lastly, click on <FileName>.ipynb.
• To run on your local machine,
  1. Install Jupyter notebook using pip install jupyter (Assuming you have Python 3 already installed).
  2. In your terminal, run this command jupyter notebook. It should open a window in your browser, if it didn't then enther this address in your browser http://localhost:8889.
  3. Now navigate to <FileName>.ipynb and open it.

Resources

Theoretical

1. Course on Convolutional Neural Networks from Stanford.
2. Vizualization tool for interpreting Convolutional Neural Network responsivity

Practical

1. Youtube video by TensorFlow team: https://www.youtube.com/watch?v=5ECD8J3dvDQ
2. Amazing article on Transfer Learning: https://towardsdatascience.com/a-comprehensive-hands-on-guide-to-transfer-learning-with-real-world-applications-in-deep-learning-212bf3b2f27a