When training models, we should place data and code in the same location as much as possible.

Keeping them in the same location helps avoid path-related issues, such as needing to specify absolute paths for the data.

For example, if I set the path to ./data/, I only need to place the data in the ./data directory.

I can do this by:

1

ln -s $(source-path-of-dataset) ./data

To link data from other locations.

If on the same host, git can automatically synchronize these links.

However, if on different hosts, you need to manage them yourself.