To get the most out of this book

As mentioned earlier, you want to be very happy in Python development to absolutely maximize your time in this book. The pages don’t spend a lot of time focusing on the software, but again, everything in the GitHub repository is Python. If you’re already using a few key AWS services, like Amazon SageMaker, S3 buckets, ECR images, and FSx for Lustre, that will speed you up tremendously in applying what you’ve learned here. If you’re new to these, that’s ok, we’ll include introductions to each of these.

AWS Service or Open-source software framework	What we’re using it for
Amazon SageMaker	Studio, notebook instances, training jobs, endpoints, pipelines
S3 buckets	Storing objects and retrieving metadata
Elastic Container Registry	Storing Docker images
FSx for Lustre	Storing large-scale data for model training loops
Python	General scripting: including managing and interacting with services, importing other packages, cleaning your data, defining your model training and evaluation loops, etc
PyTorch and TensorFlow	Deep learning frameworks to define your neural networks
Hugging Face	Hub with more than 100,000 open-source pretrained models and countless extremely useful and reliable methods for NLP and increasingly CV
Pandas	Go-to library for data analysis
Docker	Open-source framework for building and managing containers

If you are using the digital version of this book, we advise you to access the code from the book’s GitHub repository (a link is available in the next section), step through the examples, and type the code yourself. Doing so will help you avoid any potential errors related to the copying and pasting of code.