Skip to content

Using fastai library to create multi-label classification model on apparel dataset

License

Notifications You must be signed in to change notification settings

kais-viz/multi-label-classification-fastai

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 

Repository files navigation

Multi-label Classification using FastAi Library

Context

Applying Fastai library on apparel image dataset and creating a multi-label classification model based on what I learned from Jeremy Howard's lesson 3 of the fastai course.

Dataset

While searching the internet for a good dataset to apply the multi-label classification on, I stumbled upon pyimagesearch's multi-label classification with keras's article, and Adrian used a small simple dataset containing 3 clothing categories. But to expand on the dataset, I combined it with trolukovich's dataset and my own by scraping Google and Bing using cwerner's fastclass package. Now it contains 8 different apparel categories in 9 different colours. It is published on Kaggle under the name Apparel Dataset

If you want to create your own image set, I highly recommend using Christian Warner's fastclass package, he explains how to use it in a short article. Additionally, there is a tutorial on pyimagesearch which helps you build an image dataset by scraping bing, but it uses a more difficult approach and requires bing API which, if you are not a student, will require you to input your credit card information along side phone verification.

For this jupyter notebook, I will be applying the fastai library to classify the apparel and its colour within an image.

About

Using fastai library to create multi-label classification model on apparel dataset

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published