PyTorchIntroduction - GitHub Pages

CSI4108-01 ARTIFICIAL INTELLIGENCE

1

PyTorch Introduction

2018. 5. 11.

Lee, Gyeongbok

PyTorch Introduction

2

Contents

? Very Basic about PyTorch

? Tensor

?

? Network Composition

?

? Some Examples

? Simple Linear Network (1x1) ? Image classification with CNN

?

PyTorch Introduction

3

Recap: Training Process

1. Sample labeled data (batch input)

2. Forward it through the network, get predictions

4. Update the network weights

3. Back-propagate the errors

Optimize (min. or max.) objective/cost function () Generate error signal that measures difference between predictions and target values



PyTorch Introduction

4

Deep Learning Frameworks

? Early days:

? Caffe, Torch, Theano

? Tensorflow (by Google) ? PyTorch (by Facebook Research) ? DyNet (by CMU)

? Keras (TF backend)

PyTorch Introduction

5

Why use framework?

Without using framework...

We needs to define all of things

Forward definition Loss definition

Backpropagate definition Manual weight update

 ................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download