XeLabs/cs-self-learning
2
0
Fork 0
mirror of https://github.com/PKUFlyingPig/cs-self-learning.git synced 2026-06-26 11:26:58 +08:00
Code Issues Projects Releases Packages Wiki Activity
cs-self-learning/docs/机器学习系统/AICS.en.md

3.3 KiB
Raw Blame History

Intelligent Computing Systems

Course Overview

  • University: University of Chinese Academy of Sciences
  • Prerequisites: Computer Architecture, Deep Learning
  • Programming Languages: Python, C++, BCL
  • Course Difficulty: 🌟🌟🌟
  • Estimated Hours: 100+ hours

Intelligent computing systems serve as the backbone for global AI, producing billions of devices annually, including smartphones, servers, and wearables. Training professionals for these systems is critical for China's AI industry competitiveness. Understanding intelligent computing systems is vital for computer science students, shaping their core skills.

Prof. Yunji Chen's course, taught in various universities, uses experiments to provide a holistic view of the AI tech stack. Covering deep learning frameworks, coding in low-level languages, and hardware design, the course fosters a systematic approach.

Personally, completing experiments 2-5 enhanced my grasp of deep learning frameworks. The BCL language experiment in chapter five is reminiscent of CUDA for those familiar.

I recommend the textbook for a comprehensive tech stack understanding. Deep learning-savvy students can start from chapter five to delve into deep learning framework internals.

Inspired by the course, I developed a simple deep learning framework and plan a tutorial. Written in Python, it's code-light, suitable for students with some foundation. Future plans include more operators and potential porting to C++ for balanced performance and efficiency.

Course Resources

  • Course Website：Official Website
  • Course Videos：bilibili
  • Course Textbook："Intelligent Computing Systems" by Chen Yunji

Course Experiments and Resources

New Edition Experiments for 2024

  • The 2024 edition of the Intelligent Computing Systems lab has undergone extensive adjustments in the knowledge structure, experimental topics, and lab manuals, including comprehensive use of PyTorch instead of TensorFlow, and the addition of experiments related to large models.
  • As the new lab topics and manuals have not been updated on the Cambricon Forum, the following repository is provided to store the new versions of the Intelligent Computing Systems lab topics, manuals, and individual experiment answers:
  • The resources for the new edition will be updated following the course schedule of the UCAS Spring Semester 2024, with completion expected by June 2024.
  • New edition labs, manuals, and answers: https://github.com/Qudoudou-1001/2024-AICS-EXP

Old Edition Experiments