Max Balandat

We've been contemplating using pyre to do type checking and we'd really like to do that, unfortunately in the past this has caused a lot of headaches due to pyre...

> So how about a more pragmatic approach? Instead of fixing pytorch issues, you could simply ignore them entirely (or in whatever reasonable way) but still make your own type...

@AdrianSosic as of #2982 we do have a `py.typed` file - we enabled type checking in botorch itself but will look into how painful this becomes :)

This would be nice to have - would you be willing to put together a design / draft implementation for this?

This is a very high parallelism setting, the Hypervolume-based multi-objective algorithms we use by default in the `BOTORCH_MODULAR` strategy are somewhat computationally intensive, so I'm not surprised this takes a...

> would not a for loop for suggestions (aka get_trial) will result in the same suggestion over and over? if not, why? Isn't it just looking for max on the...

@chanansh which version of Ax are you referring to? The current version has an updated API and doesn't define a `get_next_trial()` method (only a `get_next_trials()` one).

> Is MultiTaskGP appropriate for multi-output data like this? My understanding is now that tasks ≠ metrics. But I want to confirm whether MultiTaskGP is intended for: multi-output regression or...

cc @saitcakmak who has done some investigations in the past into removing LinearOperator overhead. I'll let him give the details, but IIRC the high level tl;dr was that while we...

I am not sure if there were any special considerations at the time - @gpleiss or @jacobrgardner might know?