Kevin Greenman

We are closing all enhancement requests for chemprop v1.x, as we will not have time to implement these before the upcoming release of [chemprop v2.0](https://github.com/chemprop/chemprop/issues/517). If there is still interest...

This was removed in #695 since the file was out of date, but we should consider re-adding an updated version if we can get it to work. This was requested...

Is there any update on plans to move these to a descriptastorus namespace or on the new descriptors being implemented?

Found this issue while responding to a similar [question](https://github.com/chemprop/chemprop/issues/298) on the chemprop repo. Hopefully that helps if you were still wondering this.

thanks for the PR! It looks like the tests that are failing are Mac OS (has been failing for a while due to the torch-scatter installation issue) and Python 3.12,...

@JacksonBurns I'm okay with the warning option. I don't think we should change the inheritance.

Does this still work without the chemprop commit hash specified, as long as the torch and torch scatter versions are specified? If it works on the latest commit, is there...

note that the default `seed` of `MolGraphDataLoader` is still `None` and was not changed by #684

@c-w-feldmann thanks for the suggestion, I assume you're referring to [this](https://github.com/scikit-learn/scikit-learn/blob/main/sklearn/base.py#L178-L200)?

I think our conclusion in #469 was that it probably wasn't necessary but wouldn't hurt, so we included it. Perhaps we should include it for consistency, but I don't think...