Hello everyone! This notebook is my submission to the Digit Recognizer competition from Kaggle and also a quick introduction to FastAI and PyTorch. The task was to create a machine learning model that can identify hand-written digits. For this purpose, the MNIST dataset was provided.
PyTorch is one of the best Python libraries used for developing deep learning systems. FastAI just adds some high-level functionalities on top of PyTorch. In this notebook, I am going to provide some context into how a neural network can be easily implemented using these tools. Moreover, every line of code from this notebook is going to be explained to provide more insights into how you can apply these notions yourself. That being said, let's get started and good luck with your code!
If you want to find out more about the capabilities of the FastAI library, I strongly recommend checking out https://course.fast.ai/?ref=blog.paperspace.com, a course designed by the co-founder of fast.ai, Jeremy Howard.
Here are the official documentations for these two libraries:
- PyTorch: https://pytorch.org/docs/stable/index.html
- FastAI: https://docs.fast.ai/