D-X-Y / AutoDL-Projects

Automated deep learning algorithms implemented in PyTorch.
MIT License
1.56k stars 281 forks source link

Where is the documentation #126

Open gasasaaa opened 1 year ago

gasasaaa commented 1 year ago

Hello ! I'm sorry if I'm only here to complain, but I've struggled so much to install correctly dependencies for this project, with xautodl not being compatible with latest numpy version (hence latest python version / torch version), the PYTORCH_path not set up correctly etc...

And now that I have everything set, I just don't understand anything about project structure. I'd like to run a few tests with DARTS, which is the basic of NAS, as I'm currently learning. I don't understand where it is implemented, the repository is just full of .sh files doing the same thing with various parameters. The basic command given in the DARTS examples of the readme refers to files that don't exist anymore, and I keep getting random error when trying to execute the other scripts.

Could it be possible to have a basic example.py file to execute, to check that everything is set properly ? Then I'll deep dive through the code, but without seeing it work at least on some basic examples, it's just discouraging. And sure, you can't describe every line of your code, but could we at least have an overview of your project structure and how you organized your code ? I don't consider myself a total novice, and I've already spent several days trying to make everything work.