Deep learning workshop notebooks

Work-in-progress notebooks, to be used in a deep learning workshop I'm putting together.

Planned outline:

Neural network intro

• 01-regression-keras.ipynb -- basic linear regression with toy data
• 02 - MLP, notmnist, keras -- one layer, then 2 or 3.
• 03 - monitoring performance, validation sets, etc -- illustrate what's going on above. Keras.
• 03.5 - Tuning -- dropout. Weight penalties.
• 04 - CPU GPU. Keras.
• 05 - TF demo. Same dataset.

CNNs

• 01 - intro -- Keras, notmnist
• 02 - data augmentation demo

Text, RNN, LSTM

• TBD

RL

• TBD

Autoencoders

• TBD

Tuning and exploration

• 01 - cars

Datasets

• 01-notmnist-setup -- explore dataset, create pickled file for later use (based on TF example)