CPSC 330 Lecture 13: neural networks & computer vision

preview_player
Показать описание

0:00 course roadmap
2:20 multi-class classification
21:40 computer vision intro
23:40 neural networks intro
32:50 deep learning software
35:40 neural networks for images
45:55 pre-trained networks
57:40 transfer learning
1:09:40 fun stuff
1:11:30 summary
  1. Mike Gelbart
Рекомендации по теме
CPSC 330 Lecture 1:12:17

CPSC 330 Lecture 13: neural networks & computer vision

CPSC 330 Lecture 1:05:10

CPSC 330 Lecture 15: natural language processing

CPSC 330 Lecture 1:13:12

CPSC 330 Lecture 7: imputation, scaling numeric features, `ColumnTransformer`

CPSC 330 Lecture 1:13:23

CPSC 330 Lecture 14: distances and neighbours

Unsupervised Learning of 0:05:07

Unsupervised Learning of Probably Symmetric Deformable 3D Objects from Images in the Wild

CPSC 330 Lecture 0:38:21

CPSC 330 Lecture 22: ethics & responsible use

'Calling A Girl 0:03:59

'Calling A Girl Was Our Vietnam!' - Michael Gelbart

Complete Roadmap for 0:01:01

Complete Roadmap for Deep Learning | Learn Deep Learning for Free | Satyajit Pattnaik

CPSC 330 Lecture 1:06:27

CPSC 330 Lecture 16: outliers

Improving Graph Neural 0:12:37

Improving Graph Neural Networks with Structural Adaptive Receptive Fields

CPSC 330 Lecture 1:04:22

CPSC 330 Lecture 5: pipelines & hyperparameter optimization

CPSC 330 Lecture 0:51:18

CPSC 330 Lecture 23: model deployment & conclusion

Machine Learning: Linear 0:08:31

Machine Learning: Linear Regression | How to evaluate the model, error metrics with examples

Neural Networks Lecture 0:30:47

Neural Networks Lecture 1

[DSCI 511] Functions 1:14:24

[DSCI 511] Functions in Python

D012554 UGent -- 0:24:49

D012554 UGent -- 0. Introduction

CPSC 330 Lecture 1:14:19

CPSC 330 Lecture 17: time series data

Bigger batch size 0:02:47

Bigger batch size reduce training time

Neural network course 0:08:59

Neural network course - Backpropagation Output Layer [7]

Graph-Informed Neural Networks 0:19:04

Graph-Informed Neural Networks by Søren Taverniers

[CPSC 340] Clustering 0:47:23

[CPSC 340] Clustering

Lecture 10_1 | 0:34:10

Lecture 10_1 | Linear Regression | Artificial Neural Networks

IBA: Intro to 0:31:52

IBA: Intro to AI - Lecture 19 - Machine Learning - Artificial Neural Networks(3)

DLRL Summer School 1:00:59

DLRL Summer School 2021 - Online Hyperparameter Tuning - Kyunghyun Cho

Комментарии
Автор

Correction: in the dog breed classification demo I accidentally violated the "Golden Rule" - turns out the dog breed dataset is a *subset* of the ImageNet dataset, which the pre-trained network was trained on. Therefore, some (or most?) of the dog breed images in the validation/test sets were likely in the training set when the original network was trained. Whoops. Furthermore, this is not really a realistic application of transfer learning to simply apply the model to a subset of the original dataset, and the performance is unreasonably good as a result. I did not realize this at first!

mgelbart