faster.ai

• Day 1:

  1. (slides) introductory slides
  2. (code) a first example on Colab: dogs and cats with VGG
  3. (code) making a regression with autograd: intro to pytorch
  4. using colab to compute features, just run the following notebook on colab. It should load the features of dogs and cats in your drive.

• Day 2:

  1. (slides) refresher: linear/logistic regressions, classification and PyTorch module.
  2. using colab features to overfit . This practical requires the features computed on colab in day 1.
  3. (code) understanding convolutions and your first neural network for a digit recognizer.
  4. (slides) embeddings and dataloader
  5. (code) Collaborative filtering: matrix factorization and recommender system

• Day 3:

  1. (slides) Reccurrent Neural Networks and associated code
  2. (code) PyTorch tutorial on char-RNN
  3. (slides) optimization
  4. (code) Word2vec
  5. (code) Playing with word embedding

• Day 4:

  1. (slides) Generative Adversarial Networks
  2. (code) Conditional and Info GANs
  3. (slides) Opening the black box
  4. (code) CAM
  5. (code) Adversarial examples

Back to dataflowr