Popular machine learning libraries and platforms
What an exciting time to be alive
Machine learning is a rapidly growing field with the world's best minds pushing its development
New toolkits and libraries are coming out so fast making data scientists' life exponentially easier and it's also ever so hard to keep up
Here I have a few of them that I really liked: PyTorch, TensorFlow, FastAI(built upon PyTorch and its my personal favourite, you will see why)
The reason that I make them into repos is that I refer back to them from time to time when building data sicence pipelines for new projects
Hope these useful tricks and scripts could help your projects as well ^ ^
PyTorch
Definetly a library that is gaining more popularity in both academic and industry
Definetly not dont miss out this one
Here I will share 145 PyTorch tricks and some advanced techniques within the library that will make your life easier
FastAI
This is so far my favourite library among all deep learning libraries
It's super flexible to customize, has a good collection of models and a super nice touch with the callbacks
Oh, the best thing is: by exploring the source code you will learn tons of the state-of-the-art deep learning techniques
Productionize DL solutions on GCP with TensorFlow
When it comes to productionization and distributed training you can't go wrong with TensorFlow
Here we have detailed steps and code for developing and deployment of your DL solutions on Google Could Platform(GCP)
GCP Cloud Vision
Quick implementation of could vision models with AutoML Vision for everyone as well as Cloud Vision API for more experienced data scientists
GCP time series and NLP
A quick start with pre-trained BLP models with AutoML and end-to-end developer platform with DialogFlow
GCP Recommendation Systems
Intro and implementation to collaborative filtering, Knowledge-based, content-based recommenders as well as hybrid systems with TensorFlow.
Swiss knife approach
A good data scientist is like a swiss knife who has wide range of capabilities in different machine learning applications.
It's imperitive to build solid a knowledge foundation across a myriad of ML fields so that we can swiftly adapt to this rapidly evolving industry.