This is a tutorial for deep learning course which includes some experiments. Every experiment uses tensorflow as the implementation platform.

tf-model-manip.py

• A tool for checking your network design.
• Useage: tf-model-manip.py model.py
• Requirements:
  1. a python file called model.py
  2. a class called Model with a build method in model.py

Docker-materials

• A folder includes some files to help you create a docker for experiments.

Note

• If this is your first time doing homeworks, please create a docker first. You can read Docker.md for more informations.

Course slides

https://github.com/zsc/megvii-pku-dl-course/tree/master/slides19

GPU servers

• We provide some servers, each server has eight Nvidia GPU for students to do all the experiments.
• python and tensorflow has already been installed.
• Some lucky guys will get a IP address and a port number to log in a docker on servers.
• You can read Docker.md for more information.

Warnings

• Do not use the computation resource for your private project!!
• Do not let anyone who has not selected this course know the username and password of your docker!!