d2l-ai / 1day-notebooks

d2l-1day-notebooks

Notebooks for a 1-day crash course. It aims for teaching deep learning in a single day. This repo contains the notebooks with only simplified code blocks. The texts are also summarized into slides that will be uploaded later.

Check the wiki page for instructions to setup the running environments.

Part 1: Deep Learning Basic

title	ipynb	slides
Data Manipulation with Ndarray	github	nbviewer
Automatic Differentiation	github	nbviewer
Linear Regression Implementation from Scratch	github	nbviewer
Concise Implementation of Linear Regression	github	nbviewer
Image Classification Data (Fashion-MNIST)	github	nbviewer
Implementation of Softmax Regression from Scratch	github	nbviewer
Concise Implementation of Softmax Regression	github	nbviewer
Implementation of Multilayer Perceptron from Scratch	github	nbviewer
Concise Implementation of Multilayer Perceptron	github	nbviewer

Part 2: Convolutional Neural Networks

title	ipynb	slides
GPUs	github	nbviewer
Convolutions	github	nbviewer
Pooling	github	nbviewer
Convolutional Neural Networks (LeNet)	github	nbviewer
Deep Convolutional Neural Networks (AlexNet)	github	nbviewer
Networks Using Blocks (VGG)	github	nbviewer
Inception Networks (GoogLeNet)	github	nbviewer
Residual Networks (ResNet)	github	nbviewer

Part 3: Performance

title	ipynb	slides
A Hybrid of Imperative and Symbolic Programming	github	nbviewer
Multi-GPU Computation Implementation from Scratch	github	nbviewer
Concise Implementation of Multi-GPU Computation	github	nbviewer
Fine Tuning	github	nbviewer

Part 4: Recurrent Neural Networks

title	ipynb	slides
Text Preprocessing	github	nbviewer
Implementation of Recurrent Neural Networks from Scratch	github	nbviewer
Concise Implementation of Recurrent Neural Networks	github	nbviewer
Gated Recurrent Units (GRU)	github	nbviewer
Long Short Term Memory (LSTM)	github	nbviewer