# Can I restore a Tune run without checkpointing the Trainable?

**URL:** <https://discuss.ray.io/t/can-i-restore-a-tune-run-without-checkpointing-the-trainable/7559>\
**Category:** Uncategorized\
**Created:** [September 14, 2022, 3:52pm UTC](https://discuss.ray.io/t/can-i-restore-a-tune-run-without-checkpointing-the-trainable/7559 "2022-09-14T15:52:32Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![mgerstgrasser](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/mgerstgrasser/32/2728_2.png) [@mgerstgrasser](https://discuss.ray.io/u/mgerstgrasser)\
**Post date:** [September 14, 2022, 3:52pm UTC](https://discuss.ray.io/t/can-i-restore-a-tune-run-without-checkpointing-the-trainable/7559/1 "2022-09-14T15:52:32Z")

</div>

If I have a trainable function that doesn’t do any checkpointing, can I still use `Tuner.restore()`? I just want the HyperOpt search algorithm to resume from where it left of, but it’s fine to restart (rather than resume) any trials that were in progress when the previous run was interrupted.