# Reporting validation auc when hyperparameter training with raytune

**URL:** <https://discuss.ray.io/t/reporting-validation-auc-when-hyperparameter-training-with-raytune/2755>\
**Category:** Ray Tune\
**Created:** [July 2, 2021, 4:45pm UTC](https://discuss.ray.io/t/reporting-validation-auc-when-hyperparameter-training-with-raytune/2755 "2021-07-02T16:45:38Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![gagasth](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/gagasth/32/812_2.png) [@gagasth](https://discuss.ray.io/u/gagasth)\
**Post date:** [July 2, 2021, 4:45pm UTC](https://discuss.ray.io/t/reporting-validation-auc-when-hyperparameter-training-with-raytune/2755/1 "2021-07-02T16:45:39Z")

</div>

Has anyone used raytune with keras? I am trying to output a custom metrics: validation auc and difference between train and test AUC. Is there a way to do this? I haven’t been able to figure this out from the available examples on raytune. Thank you.