Shoutout to the awesome things fastai library does


(Gokul) #1

Hacker-earth conducted an image classification problem to classify the type of shirts of http://www.myntra.com/
Was able to crack top 10. As I worked with the dataset and the approach, was amazed at most of the things fastai library does. Abstracting away the architectures itself, augmentation of images, test time augmentation, learning finder etc etc

Thanks to the awesome work of fastai libray and shout out to the people involved. It really is helpful

Note: I had taken part 1 course without the fastai library, and things were a bit difficult than it had to be.
Wrote a small blog about the problem: https://medium.com/@Stormblessed/diving-into-deep-learning-part-i-ensembling-to-build-world-class-image-classifier-305e93fa8e20