Miles Olson comments

Results 30 comments of


                                            Miles Olson

Parameter type checking

Thank you for bringing this to our attention, and I'm happy you were able to unblock yourself! What you're suggesting does not sound unreasonable to me; numpy (and torch) floats...

A WARNING from ax.storage.sqa_store.save when saving an Experiment with MapData

Alex, we're having some trouble reproducing this error on our end. Could we ask you for a minimal reproduction of this error and a full stack trace?

Repeated trials without convergence

Thank you for reaching out about this. I have a good feeling your hunch about the acquisition function optimization could be on the right track -- @danielrjiang @SebastianAment could you...

Error in using Botorch models

Its hard to say exactly what the root cause of this is -- could you provide a minimal repro for us to take a look at? There could be a...

`AxClient.get_best_trial()` produces the wrong best trial index with `use_model_predictions=True`.

Thanks for catching this, this is indeed a bug on our end. @saitcakmak could you take a look at fixing this?

API docs still show usage of minimize=True and objective_name="branin" despite deprecation warning

Thanks for the heads up, let me put up a PR right now to fix that. Also, this tutorial for the Service API specifically might be a better resource (and...

[Question] Number of trials and batches in offline optimization + tracking issue to learn of experiment outcomes

Hi Eric thank you for reaching out. The number of trials Ax needs to converge on an optimal parameterization can very depending on the specifics of the experiment. In general...

Submitit ask-tell tutorial

Looks great! A couple notes (that I will just leave here since commenting on a .ipynb is annoying): * I can handle making sure this renders properly on the website....

Submitit ask-tell tutorial

One last thing: no need to write the covariance since the demo problem only has one metric

Idea for interactive template generation: multi-objective, multi-fidelity, batch optimization, etc.

This is a neat idea! We are always looking for ways to improve our documentation and tutorials to make learning Ax easier. We already have some plans for improving docs...