https://community.databricks.com/t5/machine-learning/what-s-a-best-practice-for-hyperopt-workflow/m-p/21434#M1172 <UL><LI>Choose what hyperparameters are reasonable to optimize</LI><LI>Define broad ranges for each of the hyperparameters (including the default where applicable)</LI><LI>Run a small number of trials</LI><LI>Observe the results in an MLflow parallel coordinate plot and select the runs with lowest loss</LI><LI>Move the range towards those higher/lower values when the best runs’ hyperparameter values are pushed against one end of a range</LI><LI>Determine whether certain hyperparameter values cause fitting to take a long time (and avoid those values)</LI><LI>Re-run with more trials</LI><LI>Repeat until the best runs are comfortably within the given search bounds and none are taking excessive time</LI></UL><P><A href="https://databricks.com/blog/2021/04/15/how-not-to-tune-your-model-with-hyperopt.html" target="test_blank">https://databricks.com/blog/2021/04/15/how-not-to-tune-your-model-with-hyperopt.html</A></P> Wed, 23 Jun 2021 14:41:19 GMT User16789201666 2021-06-23T14:41:19Z https://community.databricks.com/t5/machine-learning/what-s-a-best-practice-for-hyperopt-workflow/m-p/21434#M1172 <UL><LI>Choose what hyperparameters are reasonable to optimize</LI><LI>Define broad ranges for each of the hyperparameters (including the default where applicable)</LI><LI>Run a small number of trials</LI><LI>Observe the results in an MLflow parallel coordinate plot and select the runs with lowest loss</LI><LI>Move the range towards those higher/lower values when the best runs’ hyperparameter values are pushed against one end of a range</LI><LI>Determine whether certain hyperparameter values cause fitting to take a long time (and avoid those values)</LI><LI>Re-run with more trials</LI><LI>Repeat until the best runs are comfortably within the given search bounds and none are taking excessive time</LI></UL><P><A href="https://databricks.com/blog/2021/04/15/how-not-to-tune-your-model-with-hyperopt.html" target="test_blank">https://databricks.com/blog/2021/04/15/how-not-to-tune-your-model-with-hyperopt.html</A></P> Wed, 23 Jun 2021 14:41:19 GMT https://community.databricks.com/t5/machine-learning/what-s-a-best-practice-for-hyperopt-workflow/m-p/21434#M1172 User16789201666 2021-06-23T14:41:19Z