Deep Learning with PyTorch for beginners

A quick context: I’m not a deep learning expert but I’m an aspirant. I have been acquiring data science skills through several means — online training, reading blogs, practicing python libraries, decoding some of the Kaggle experiment submissions, etc. Some of what I learn stays with me and some don’t. Still, every time I try something new, I make connections with my past learning. That gives me the determination to pursue learning despite difficulties. After more than a year of relentless pursuing, I can connect the dots very well now.

The online course Deep Learning with PyTorch: Zero to GAN offered by Jovian is one such learning opportunity that I wanted to make use of. Though it’s a bit of stretch given my current expertise on machine learning. I love to jump a few steps far ahead to know what lies there. It is interesting so far and it will be.

If you see yourself in a similar situation as I’m in, feel free to jump into this course and learn. It’s a free course but the quality of teaching, assignment, and collaboration are of a high standard.

Let’s get into the subject.

PyTorch is an open-source machine learning library mainly used for applications such as Computer Vision and Natural language processing (NLP). This was developed by Facebook’s AI research lab.

Watch this first lecture for initial setup and get started with PyTorch: link

This blog is based on the first assignment from the course which is to explore 5 interesting functions from TORCH.TENSOR library documentation. The detailed assignment can be found on this link. Here I’ll give a glimpse of what it is all about.

Tensor is a basic element of the Torch library. A tensor is a number, vector, matrix, or any n-dimensional array. PyTorch is a library for processing tensors. Here is how you create a tensor by feeding in an array of values:

Here are the key functions of PyTorch explained in detail:

I hope you found this helpful.

Please leave a comment if any queries or feedback.