Oxford 102 Flower Classification

January 30, 2019

Trained an image classifier to recognize different species of flowers, and build web app with Flask. The flowers chosen to be flower commonly occuring in the United Kingdom. Dataset can be found at Kaggle ’s InClass competitions.

Sezin Engur

Engineer | ML Enthusiast

Python Developer

Istanbul

no post found

A/B Testing with a Mobile Game

Mar 3, 2019

A/B tests shows that retention is higher when the gate is positioned earlier. The data tells us about the theory of hedonic adaptation is tendency to get less and less enjoyment out of a fun activity over time if that activity is undertaken continuously. That’s why when the gate is moved from level 30 to level 40, fewer players make it far enough, and because of boredom they are more likely to quit.

Character level RNN with Netflix 's Sex Education

Feb 2, 2019

The network will train character by character on some text-Netflix ’s Sex Education season 1 scripts, then generate new characters. I trained the characters with using scripts from season 1, then get top 50 characters for each person. That’s why the sentences are in speech bubbles seems weird.

Style Transfer with PyTorch, Elon and Picasso

Feb 2, 2019

Recreate a style transfer-uses 19 layer VGG Network- method to mix photo of ‘high’ Elon Musk and Picasso ’s painting, The Dream. The method that is outlined in the paper, Image Style Transfer Using Convolutional Neural Networks, by Gatys.

Oxford 102 Flower Classification

Jan 1, 2019