Notebooks List¶
Here we have a list of example notebooks using Torchbearer with a brief description of the contents and broken down by broad subject.
General¶
Quickstart Guide:
This guide will give a quick intro to training PyTorch models with Torchbearer.
Callbacks Guide:
This guide will give an introduction to using callbacks with Torchbearer.
Imaging Guide:
This guide will give an introduction to using the imaging sub-package with Torchbearer.
Serialization:
This guide gives an introduction to serializing and restarting training in Torchbearer.
History and Replay:
This guide gives an introduction to the history returned by a trial and the ability to replay training.
Custom Data Loaders:
This guide gives an introduction on how to run custom data loaders in Torchbearer.
Data Parallel with Torchbearer:
This guide gives a brief introduction on how to use PyTorch DataParallel with Torchbearer models.
LiveLossPlot with Torchbearer:
This guide shows how we can get live loss visualisations in notebooks with LiveLossPlot.
PyCM with Torchbearer:
This guide shows how we can generate confusion matrices with PyCM in torchbearer.
Nvidia Apex with Torchbearer:
This guide shows how we can do half and mixed precision training in torchbearer.
Deep Learning¶
Training a VAE:
This guide covers training a variational auto-encoder (VAE) in Torchbearer, taking advantage of the persistent state.
Training a GAN:
This guide will cover how to train a Generative Adversarial Network (GAN) in Torchbearer using custom closures to allow for the more complicated training loop.
Class Appearance Model:
In this example we will demonstrate the ClassAppearanceModel callback included in torchbearer. This implements one of the most simple (and therefore not always the most successful) deep visualisation techniques, discussed in the paper Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps
Adversarial Example Generation:
This guide will cover how to perform a simple adversarial attack in Torchbearer.
Transfer Learning:
This guide will cover how to perform transfer learning of a model with Torchbearer.
Regularising Models:
This guide will cover how to use Torchbearers built-in regularisers.
Differentiable Programming¶
Optimising Functions:
This guide will briefly show how we can do function optimisation using Torchbearer.
Linear SVM:
This guide will train a linear support vector machine (SVM) using Torchbearer.
Breaking ADAM:
This guide uses Torchbearer to implement On the Convergence of Adam and Beyond, one of the top papers at ICLR 2018, which demonstrated a case where ADAM does not converge.