Havilah - intro

I'll get straight to the point.

I have 2 months and 1 week till uni resumes and I have to do awesome stuff before then

Here's me documenting those potentially awesome stuff ✨

By the way, I'm Bams. I am a Computer Engineering major, and I do Machine Learning, Deep Learning, some web development, and some UI. But for this period (let's call it Havilah), I'll be focusing on Deep Learning (Computer Vision) and JavaScript.

Why you ask? Cos Computer Vision is what I want to specialize in and I plan on learning react native, for mobile apps so I have to know JavaScript (I want to know how to put ML on apps)

Today I learnt how to load images for multiclass classification from their various class folders for computer vision tasks. Also how to process them to a form that the computer can work with (numbers✨)

I need to know how to do this for literally any computer vision task and later on I plan on making a classifier for hands ✋🏻🤚🏽 and for this I'll need to collect data myself (more on that later)

After I figured this out, I made a classifier for the data in Tensorflow. The dataset was a collection of images of dog breeds 🐕🐩🐕‍🦺.

I wanted to see how the model would perform if it were a plain neural network (spoiler it did terribly) so I tried it on that one first before using a Convolutional Neural Network (CNN)

A challenge I faced was the training process plateaued repeatedly, it repeatedly gave an accuracy of 0.1134 (I almost cried, almost 🌚) but it turned out I just kept fitting the same model I used for the plain neural network on the data thinking I was using the model for the CNN... yeah