Marc Romeyn

This looks awesome! Am I correct to say that these new rules would work with remote execution? And do you have an example by any chance of the setup you...

+1 on this one! Seems like a very interesting project indeed. What's your roadmap with this? Are you accepting PR's?

This seems like a very trivial PR or am I missing something @benfred @jperez999? This PR is blocking a bunch of other PRs we’d like to get in for the...

I think it could be useful to add some better docs around this method since it might be slightly confusing for users. This because we select branches by features and...

We do, see [source](https://github.com/NVIDIA-Merlin/models/blob/main/merlin/datasets/entertainment/movielens/dataset.py#L79-L82). The issue might be that TF is loaded and therefore takes up 50% GPU memory. It sounded like Oliver has a GPU with 6GB of memory,,...

That’s how we did it at Spotify as well, we presented that system [here](https://www.nvidia.com/en-us/on-demand/session/gtcspring21-s32017/). It would be very straightforward to replicate that here.

rerun tests

I am confused about why that transform needs to be in NVT? I think we have a Block in Merlin-models that does the continuous-power [source](https://github.com/NVIDIA-Merlin/models/blob/ecdd78574bd1584acdb721ecd20b1a9d1236a0e7/merlin/models/tf/blocks/core/transformations.py#L253-L280). Or is this not what...

Ranking ticket is here: https://github.com/NVIDIA-Merlin/Merlin/issues/1044

> > PL does this like [this](https://github.com/Lightning-AI/pytorch-lightning/pull/19404/files#diff-72f5c7707bc8f4033f0e5f644a8b2eda5d333084e91dfeaaa1368dcbdf24df02). So I feel like we should adopt the same design and not add a custom arg to our strategy. > > If only...