Adam J. Stewart

I think I found a few bugs. The `__init__.py` in the root of the repo breaks things for me. Also, the lack of `__init__.py` in the `tests` directories and subdirectories...

Personally, I would move all of `llnl` into `spack`. For `external`, we should try to directly use pip-installed dependencies. If there is anything in `external` that isn't on PyPI, I...

This seems to come up a lot. Especially in our land cover datasets, you almost always want to be able to select a subset of classes and then map them...

> I think you can simply do something like: Correct, but we could formalize this in a more user-friendly transform. > Why would you only want to re-map a subset...

Why can't we support both numpy 1 and 2 as soon as possible? Why do we need to wait for other dependencies? Seems like it's a simple fix.

It's not possible for your tests to pass without waiting for torchvision, but it is possible for the next release to be compatible with numpy 2 (futureproof). This way, as...

https://github.com/lightly-ai/lightly/pull/1607

Now we just need a new release 😄

No rush, lightly's release schedule is already faster than most. This is partly why I was so adamant about supporting numpy 2 immediately and letting torchvision deal with complaints about...

Unfortunately torch/torchvision still don't support numpy 2 on Windows in the latest release, but at least we can use this on macOS/Linux.