[DOC] Add cuML Pipeline Documentation
Report needed documentation
Report needed documentation There is no documentation for the cuML Pipeline object wrapper.
Describe the documentation you'd like It seems cuML provides a thin wrapper around the sklearn Pipeline object. Typically I'd expect this to require relatively little documentation, however the Pipeline object has some important default behaviors and it is not clear how they interact with cuML and GPU usage. It would be nice to have some general guidance here. For example
- The Pipeline object has a memory argument that specifies the location of temporary cached information. Can this effect throughput when using it with cuML since I assume it's not possible to configure an in memory GPU cache?
Steps taken to search for needed documentation The existence of the cuML Pipeline object is only clear through an old issue #1321 it doesn't exist in stable or nightly docs.
Thanks for the issue @ZeroCool2u, apologies for my delayed response. Documentation improvements are very high in the priority list of issues for the team, so we will look into it for the upcoming versions.