Linear Regression with PyTorch
About me
My name is Abubakar, nickname Abu 
. I recently graduated from university and hold a degree in CS (computer science). I am very passionate about AI/Machine learning and recently developed an interest in distributed systems. I work part-time as full-stack developer. I am self taught and pretty much learn every day.
Aside from that, I work out and stay active. I believe in staying busy and learning or experiencing new things is what makes us grow 
. I spend weekends with family or friends or learning to play piano. My famous quote is by Alan Turning, "Sometimes it is the people no one imagine anything of who do the things that no one can imagine."
The problem I wanted to solve
My motivation was to learn a library called PyTorch. Starting with just simple concept/task of building Linear Regression model. Once I am comfortable with the library I can build advance models of Neural Networks.
What is Linear Regression with PyTorch?
The explanation is given in that link. (Didn't want to rewrite the whole thing here)
Linear Regression
Tech stack
I choose Python 3.x, Jupyter Notebook, mat-plot and PyTorch. These are emerging technologies in field of AI.
The process of building Linear Regression with PyTorch
I got started by reading the PyTorch documentation.
Process:
Step 1: I build a small script just using Python 3.x.
Step 2: I modified the script to work in Jupyter Notebook.
Step 3: I wrote comments and explanations.
Challenges I faced
Nothing to difficult as this is very simple concept. Probably just setting up the Jupyter Notebook required just small configuration changes. Other than that nothing challenging.
Key learnings
I would set up the Jupyter Notebook from the beginning. It makes the life easier and provides much flexibility for writing good Python tutorials or just learning.
Tips and advice
I say before getting started with PyTorch, just make sure to read documentation. That will save time in future and help speed up the process.
Final thoughts and next steps
I will continue enhancing this project by using actual data from Kaggle or some other data-set websites. Then eventually moved to building Neural Networks models PyTorch.
