Skip to content

hands-on

Directory actions

More options

Directory actions

More options

Latest commit

 

History

History

hands-on

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
data
data
 
 
optional
optional
 
 
010_underfitting_overfitting_complete.ipynb
010_underfitting_overfitting_complete.ipynb
 
 
010_underfitting_overfitting_courageous.ipynb
010_underfitting_overfitting_courageous.ipynb
 
 
010_underfitting_overfitting_lazy.ipynb
010_underfitting_overfitting_lazy.ipynb
 
 
020_mnist_data_exploration_complete.ipynb
020_mnist_data_exploration_complete.ipynb
 
 
020_mnist_data_exploration_courageous.ipynb
020_mnist_data_exploration_courageous.ipynb
 
 
020_mnist_data_exploration_lazy.ipynb
020_mnist_data_exploration_lazy.ipynb
 
 
030_activation_function_complete.ipynb
030_activation_function_complete.ipynb
 
 
030_activation_functions_lazy.ipynb
030_activation_functions_lazy.ipynb
 
 
040_mnist_mlp_complete.ipynb
040_mnist_mlp_complete.ipynb
 
 
040_mnist_mlp_courageous.ipynb
040_mnist_mlp_courageous.ipynb
 
 
040_mnist_mlp_lazy.ipynb
040_mnist_mlp_lazy.ipynb
 
 
050_convolution_complete.ipynb
050_convolution_complete.ipynb
 
 
050_convolution_lazy.ipynb
050_convolution_lazy.ipynb
 
 
060_mnist_cnn_complete.ipynb
060_mnist_cnn_complete.ipynb
 
 
060_mnist_cnn_courageous.ipynb
060_mnist_cnn_courageous.ipynb
 
 
060_mnist_cnn_lazy.ipynb
060_mnist_cnn_lazy.ipynb
 
 
README.md
README.md
 
 

README.md

Hands-on notebooks

The core notebooks use PyTorch for the neural-network examples. Start Jupyter Lab from this directory so that downloaded datasets and generated model files remain below hands-on:

cd hands-on
jupyter lab

All notebooks come in at least two versions:

  • lazy: all code is ready to execute and contains no saved output; and
  • complete: the reference solution, including saved output, intended for viewing on GitHub.

For some notebooks, there is a courageous version as well, which means that you will have to complete consequential parts of the code yourself.

  1. 010_underfitting_overfitting: illustrates the concepts of underfitting and and overfitting using non-linear regression.
  2. 020_mnist_data_exploration: explores the MNIST data set and PyTorch's Dataset representation.
  3. 030_activation_function: visualization of the relevant activation functions.
  4. 040_mnist_mlp: illustrates PyTorch data loaders, explicit training and evaluation loops, and construction of a classic multilayer perceptron to recognize handwritten digits.
  5. 050_convolution: illustrates convolution as used in convolutional neural networks.
  6. 060_mnist_cnn: trains and compares convolutional neural networks in PyTorch to recognize handwritten digits.

The optional directory contains the legacy Keras-based IMDB/RNN exercises. They are retained as legacy supplementary material and are not part of the PyTorch core path.