Chen Tianqi's machine learning compilation course (free)

Non advertising , I have been looking for open classes to study on weekends and evenings recently , Especially signal and system 、 Digital signal processing 、 Machine learning and other courses , To make up for the previous missing courses （ The direction of postgraduate study is graph computing, I haven't touched the signal before work 、 Voice related courses , Even machine learning doesn't involve much ）, To build your own knowledge system ​. I found many national open classes of famous teachers in Colleges and universities on the Internet , But few people study , What a pity ​. This time, I'd like to recommend the one just opened by Chen Tianqi 《 Machine learning compiler 》 Course ​https://mlc.ai/summer22-zh/.

Course introduction

​ With the application of artificial intelligence in our daily life has become more and more common , The challenge now is how to deploy the latest AI models in different production environments . The combination of model and deployment environment brings great challenges to training and reasoning deployment . Besides , At present, the landing model also puts forward more requirements , For example, reduce software dependency 、 Comprehensive model coverage 、 Accelerate with new hardware 、 Reduce memory usage , And greater scalability .

These model training and reasoning problems , Involving machine learning programming paradigm 、 Learning based search algorithm 、 Compile optimization and compute runtime . The combination of these topics produces a new theme —— Machine learning compiler , And this direction is developing continuously . In this course , Let's talk about the key elements , Systematically study the key elements of this emerging field . We will learn some core concepts , Used to represent a machine learning program 、 Automatic optimization technology , And optimizing environment dependencies in end-to-end machine learning deployment 、 Memory and performance methods .

Course audience and background requirements

This course is aimed at a wide range of users engaged in machine learning . Machine learning is a wide range of topics in practical applications , Including machine learning scientists 、 Collaboration between multiple groups such as machine learning engineers and hardware suppliers .

This course requires background knowledge in data science and machine learning ：

• be familiar with Python Language and Numpy Use ;

• Some in-depth learning framework background knowledge （ for example PyTorch, TensorFlow, JAX）;

• System level programming experience is preferred （ for example C/CUDA）.

  ---

• 

•