In this article, I will highlight some useful packages for ML/AI practitioners from A to Z. It is also a useful starting point for Python programmer learners to sharpen their practical skills after learning the language itself.
Python is an easy and powerful language for machine learning practitioner like me. A good style makes the code easy to read and reuse. A lengthy one comes from Google. Here, I want to share with you some tips to create troubles when others want to use your code.
Actually, I want you to avoid them.
To keep things tight, a short list followed by detailed examples in below.
PocketFlow is a model compression framework open-sourced by Tencent. After a close look, I find it should be a fairly handy tool. However, there are three caveats:
I am trying to address the above problems, and in the article, I am trying to provide a userful guidance to compress your own model.
This page provide a guidance of using TFLite to convert and deploy models.
We use LeNet-like CNN model on MNIST dataset. The workflow is general, however the performance of TF Lite model (compression, accuracy) would be different based your models and datasets.
Interactive plot of ImageNet challenge benchmark. Data is queried from papers with code.