Ax
Ax copied to clipboard
Adaptive Experimentation Platform
Summary: The stopping strategy at first call to get_next_trial() raises error since there are no compleleted trials yet to call max() function on. This fixes that. Reviewed By: danielrjiang Differential...
Summary: implements improvement based global stoppping Reviewed By: danielrjiang, Balandat Differential Revision: D34705629
Summary: Add a context manager to speed up gpytorch via their provided gpt_settings object. We utilize 3 of their contexts: fast_computations, fast_pred_samples, and fast_pred_var which you can read more about...
Summary: This implements the interface as well as a simple global stopping strategy to be used by Ax service and loop API's. Reviewed By: Balandat Differential Revision: D34691150
In running the multi-objective optimization with outcome constraints using the Service API, I tried calling **ax_client.get_pareto_optimal_parameters()**, but it fails reporting: ``` File "/moo.py", line 536, in best_parameters = ax_client.get_pareto_optimal_parameters() File...
Summary: See title. This diff isn't ready for review by any means, but is an attempt to temporarily unblock https://github.com/facebook/Ax/issues/769. Differential Revision: D33853834