# Ray tune log hyperparameters to tensorboard (solved)

**URL:** <https://discuss.ray.io/t/ray-tune-log-hyperparameters-to-tensorboard-solved/84>\
**Category:** Ray Tune\
**Created:** [November 26, 2020, 4:56pm UTC](https://discuss.ray.io/t/ray-tune-log-hyperparameters-to-tensorboard-solved/84 "2020-11-26T16:56:14Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![Ahmed\_Abbas](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/ahmed_abbas/32/48_2.png) [@Ahmed\_Abbas](https://discuss.ray.io/u/Ahmed_Abbas)\
**Post date:** [November 26, 2020, 4:56pm UTC](https://discuss.ray.io/t/ray-tune-log-hyperparameters-to-tensorboard-solved/84/1 "2020-11-26T16:56:14Z")

</div>

Just started using Ray tune and got first experiment up and running in around 3-4 hours in PyTorch. I would like to see hyper-parameters for each run of ray-tune in tensorboard. I know that tensorflow has HParams which automatically supports this but can something similar be done for PyTorch easily without modifying my code intensively?

I am thinking because ray-tune internally does know which experiment used which hyper-parameters so it should not be too difficult to expose these parameters to tensorboard logs? Or is it not?

Thanks!

---

<div class="post-metadata">

**Author:** ![Ahmed\_Abbas](https://sea2.discourse-cdn.com/flex020/user_avatar/discuss.ray.io/ahmed_abbas/32/48_2.png) [@Ahmed\_Abbas](https://discuss.ray.io/u/Ahmed_Abbas)\
**Post date:** [November 26, 2020, 9:03pm UTC](https://discuss.ray.io/t/ray-tune-log-hyperparameters-to-tensorboard-solved/84/2 "2020-11-26T21:03:31Z")

</div>

My bad, I see hparams are actually there!