Please help me address the following questions:
- Is time budget considered for each dataset or all dataset? and How to set the time budget in the AutoML blind test.
- How to calculate the time spent? I mean it includes time read file, training, testing and write the results?
Thank youPosted by: vn.sju @ Dec. 12, 2017, 9:16 a.m.
* There is a time budget for each data set, you can have access to this budget in the dataset_public.info file (this information will be also available for the blind data sets). The duration column in the leaderboard shows the overall time (over all of the data sets), however, there is a limit per each of the datasets, if the budget for a data set is surpassed, the submission will not be valid for the data set
* You can use the starting kit (sample code submission in the Files section) as a guide to estimate the time for the different stages of the process.
We hope this helps,
Posted by: hugo.jair @ Dec. 13, 2017, 7 p.m.
Thank youPosted by: vn.sju @ Dec. 14, 2017, 8:53 a.m.