Working group on topics in Machine Learning

Code examples from working group in Neural Network Workshop Series

Lead by: C. Elkhill and D. Baraghosi

Faculty: Antonio Porras, PhD

Pre-requisites

• Install anaconda, setup virtual environment (i.e., my-torch)
• Install pytorch and tensorboard to this environment
• Activate the environment and launch notebooks within environment

Week 1 Summary (Feb 15)

• Introductions, what we hope to learn
• Discussion of theory versus practical training
• Agreement on meeting time (and frequency)
• Setup new environment with pytorch installed
• Tutorial: https://pytorch.org/tutorials/beginner/blitz/cifar10_tutorial.html

Week 2 Summary (Mar 1)

• Introduction to artificial neural networks
• Started tutorial on building a neural network

Week 3 Summary (Mar 15)

• Discussed Softmax transformation of logits to obtain class probabilities
• Completed tutorial on building a network
• Next tutorial: https://pytorch.org/tutorials/beginner/basics/autogradqs_tutorial.html

Week 4 Summary (Mar 29)

• Downloaded Fashion MNIST dataset
• Setup code for training a network in 3 epochs

Week 5 (Apr 12)

• Missed this week

Week 6 Summary (Apr 26)

• Began training network for various learning rates
• Used Tensorboard to evaluate training performance