# Ray Tune **URL:** https://discuss.ray.io/c/ray-tune/5.md [Latest](https://discuss.ray.io/latest.md) · [Categories](https://discuss.ray.io/categories.md) · [Tags](https://discuss.ray.io/tags.md) --- ## [About the Ray Tune category](https://discuss.ray.io/t/about-the-ray-tune-category/47)
**Author:** [@bill-anyscale](https://discuss.ray.io/u/bill-anyscale)\ **Replies:** 0\ **Last updated:** [November 17, 2020, 12:02am UTC](https://discuss.ray.io/t/about-the-ray-tune-category/47 "2020-11-17T00:02:02Z")
For questions on hyperparameter tuning and experiment management. Use Tune to search for the best model configurations by running and scheduling many parallel trials. Tune is a Python library for experiment execution an… --- ## [Ray gets stuck on SLRUM cluster without any exception](https://discuss.ray.io/t/ray-gets-stuck-on-slrum-cluster-without-any-exception/23505)
**Author:** [@mathias](https://discuss.ray.io/u/mathias)\ **Replies:** 2\ **Last updated:** [March 19, 2026, 1:55pm UTC](https://discuss.ray.io/t/ray-gets-stuck-on-slrum-cluster-without-any-exception/23505 "2026-03-19T13:55:13Z")
1. Severity of the issue: (select one) None: I’m just curious or want clarification. Low: Annoying but doesn’t hinder my work. Medium: Significantly affects my productivity but can find a workaround. High: Comple… --- ## [\`suggest\_\*\` methods in OptunaSearch don't support keyword argument passing](https://discuss.ray.io/t/suggest-methods-in-optunasearch-dont-support-keyword-argument-passing/23479)
**Author:** [@Zentino](https://discuss.ray.io/u/Zentino)\ **Replies:** 2\ **Last updated:** [February 26, 2026, 1:11pm UTC](https://discuss.ray.io/t/suggest-methods-in-optunasearch-dont-support-keyword-argument-passing/23479 "2026-02-26T13:11:34Z")
1. Severity of the issue: (select one) Low: Annoying but doesn’t hinder my work. 2. Environment: Ray version: 2.40.0 Python version: 3.11.9 OS: CentOS 8 Cloud/Infrastructure: Docker 3. What happened vs. what you ex… --- ## [How make Ray Tune + DDP training](https://discuss.ray.io/t/how-make-ray-tune-ddp-training/23125)
**Author:** [@bouachalazhar](https://discuss.ray.io/u/bouachalazhar)\ **Replies:** 3\ **Last updated:** [January 2, 2026, 12:31pm UTC](https://discuss.ray.io/t/how-make-ray-tune-ddp-training/23125 "2026-01-02T12:31:51Z")
How can I use Ray Tune + DDP PyTorch (not PyTorch Lightning)? I don’t see an example. Please review my code if you need context; everything works, but I’m unable to verify if it’s DDP training. I test it in Kaggle/Colab;… --- ## [Unable to access best results checkpoint](https://discuss.ray.io/t/unable-to-access-best-results-checkpoint/23430)
**Author:** [@koehneiph](https://discuss.ray.io/u/koehneiph)\ **Replies:** 1\ **Last updated:** [January 2, 2026, 11:39am UTC](https://discuss.ray.io/t/unable-to-access-best-results-checkpoint/23430 "2026-01-02T11:39:14Z")
1. Severity of the issue: (select one) High: Completely blocks me. 2. Environment: Ray version: 2.53.0 Python version: Python 3.10.12 OS: Ubuntu 22.04 Cloud/Infrastructure: Docker 3. What happened vs. what you expe… --- ## [Working\_dir files not copied for Tune jobs?](https://discuss.ray.io/t/working-dir-files-not-copied-for-tune-jobs/23283)
**Author:** [@Daraan](https://discuss.ray.io/u/Daraan)\ **Replies:** 4\ **Last updated:** [December 25, 2025, 6:05am UTC](https://discuss.ray.io/t/working-dir-files-not-copied-for-tune-jobs/23283 "2025-12-25T06:05:45Z")
RuntimeEnv(working\_dir…) provides the great option to copy necessary files into my working directoy. However when using a Tuner and Trials these files are not present in the trials local path. Is that intentional, is t… --- ## [(pid=gcs\_server) and (raylet) report : Failed to establish connection to the metrics exporter agent](https://discuss.ray.io/t/pid-gcs-server-and-raylet-report-failed-to-establish-connection-to-the-metrics-exporter-agent/23367)
**Author:** [@Risiamu](https://discuss.ray.io/u/Risiamu)\ **Replies:** 1\ **Last updated:** [December 10, 2025, 9:39am UTC](https://discuss.ray.io/t/pid-gcs-server-and-raylet-report-failed-to-establish-connection-to-the-metrics-exporter-agent/23367 "2025-12-10T09:39:20Z")
1. Severity of the issue: (select one) Low: Annoying but doesn’t hinder my work. 2. Environment: Ray version: ray, version 2.52.1 Python version: Python 3.11.14 OS: Win11 Cloud/Infrastructure: None Other libs/tools … --- ## [Population Based Training Does Not Clone When Loggin Is Disabled](https://discuss.ray.io/t/population-based-training-does-not-clone-when-loggin-is-disabled/23362)
**Author:** [@knods3k](https://discuss.ray.io/u/knods3k)\ **Replies:** 1\ **Last updated:** [December 7, 2025, 10:44am UTC](https://discuss.ray.io/t/population-based-training-does-not-clone-when-loggin-is-disabled/23362 "2025-12-07T10:44:30Z")
1. Severity of the issue: (select one) Low: Annoying but doesn’t hinder my work. 2. Environment: Ray version: 2.52.0 Python version: 3.12.3 OS: macOS 15.6.1 Cloud/Infrastructure: / Other libs/tools (if relevant): to… --- ## [Repeated hyper-parameters in different runs](https://discuss.ray.io/t/repeated-hyper-parameters-in-different-runs/23347)
**Author:** [@luchino](https://discuss.ray.io/u/luchino)\ **Replies:** 1\ **Last updated:** [December 1, 2025, 5:48pm UTC](https://discuss.ray.io/t/repeated-hyper-parameters-in-different-runs/23347 "2025-12-01T17:48:14Z")
I ran a hyper-parameter tuning experiment with a few discrete parameters that get selected using tune.choice for a total of 24 total parameters. I then set the num\_samples parameter in tune.Tuneconfig precisely to 24. I… --- ## [How to avoid packaging when performing hyperparameter tuning on local machine?](https://discuss.ray.io/t/how-to-avoid-packaging-when-performing-hyperparameter-tuning-on-local-machine/23304)
**Author:** [@Jason\_Harris](https://discuss.ray.io/u/Jason_Harris)\ **Replies:** 1\ **Last updated:** [November 5, 2025, 5:27am UTC](https://discuss.ray.io/t/how-to-avoid-packaging-when-performing-hyperparameter-tuning-on-local-machine/23304 "2025-11-05T05:27:10Z")
1. Severity of the issue: (select one) High: Completely blocks me. 2. Environment: Ray version: 2.49.2 Python version: 3.12.2 OS: Mac Cloud/Infrastructure: none Other libs/tools (if relevant): none 3. What happened… --- ## [Accessing GP Surrogate and Adding Constraints in BayesOptSearch](https://discuss.ray.io/t/accessing-gp-surrogate-and-adding-constraints-in-bayesoptsearch/23277)
**Author:** [@Stan](https://discuss.ray.io/u/Stan)\ **Replies:** 1\ **Last updated:** [October 28, 2025, 10:51am UTC](https://discuss.ray.io/t/accessing-gp-surrogate-and-adding-constraints-in-bayesoptsearch/23277 "2025-10-28T10:51:34Z")
1. Severity of the issue: (select one) None: I’m just curious or want clarification. 2. Environment: Ray version: 2.50.1 I’m using BayesOptSearch for a Bayesian optimization of my simulation parameters. This works … --- ## [Limit amount of parallel trials without using max\_concurrent\_trials?](https://discuss.ray.io/t/limit-amount-of-parallel-trials-without-using-max-concurrent-trials/23244)
**Author:** [@Daraan](https://discuss.ray.io/u/Daraan)\ **Replies:** 4\ **Last updated:** [October 22, 2025, 11:21am UTC](https://discuss.ray.io/t/limit-amount-of-parallel-trials-without-using-max-concurrent-trials/23244 "2025-10-22T11:21:12Z")
TL;DR: I want to add N trials to my tuner (all have to be in PENDING/PAUSED state) and execute M in parallel until I pause them again. I work with a setup similar to population based training with forking the currently … --- ## [Multi-GPU ray + accelerate HPO](https://discuss.ray.io/t/multi-gpu-ray-accelerate-hpo/23250)
**Author:** [@pe4eniks](https://discuss.ray.io/u/pe4eniks)\ **Replies:** 1\ **Last updated:** [October 16, 2025, 5:01pm UTC](https://discuss.ray.io/t/multi-gpu-ray-accelerate-hpo/23250 "2025-10-16T17:01:32Z")
1. Severity of the issue: (select one) None: I’m just curious or want clarification. Low: Annoying but doesn’t hinder my work. Medium: Significantly affects my productivity but can find a workaround. High: Comple… --- ## [Help investigating: Wandb offline experiments sometimes not synced - Silently dead Syncer?](https://discuss.ray.io/t/help-investigating-wandb-offline-experiments-sometimes-not-synced-silently-dead-syncer/23041)
**Author:** [@Daraan](https://discuss.ray.io/u/Daraan)\ **Replies:** 0\ **Last updated:** [August 25, 2025, 8:40am UTC](https://discuss.ray.io/t/help-investigating-wandb-offline-experiments-sometimes-not-synced-silently-dead-syncer/23041 "2025-08-25T08:40:04Z")
Severity of the issue: Medium: Significantly affects my productivity but can find a workaround. I store my wandb offline and then after tuning finished want to upload them. Very roughly: results = tuner.fit() # wit… --- ## [\[RLlib,Tune\] Relevance of \_\_ref\_ph in sample\_collector experiment state](https://discuss.ray.io/t/rllib-tune-relevance-of-ref-ph-in-sample-collector-experiment-state/20813)
**Author:** [@PhilippWillms](https://discuss.ray.io/u/PhilippWillms)\ **Replies:** 1\ **Last updated:** [August 18, 2025, 10:34pm UTC](https://discuss.ray.io/t/rllib-tune-relevance-of-ref-ph-in-sample-collector-experiment-state/20813 "2025-08-18T22:34:02Z")
How severe does this issue affect your experience of using Ray? Low: It annoys or frustrates me for a moment. Medium: It contributes to significant difficulty to complete my task, but I can work around it. Dear commun… --- ## [Doc says Tune Trial Schedulers can modify hyperparamters - but how to do it?](https://discuss.ray.io/t/doc-says-tune-trial-schedulers-can-modify-hyperparamters-but-how-to-do-it/22956)
**Author:** [@Daraan](https://discuss.ray.io/u/Daraan)\ **Replies:** 2\ **Last updated:** [August 7, 2025, 10:26am UTC](https://discuss.ray.io/t/doc-says-tune-trial-schedulers-can-modify-hyperparamters-but-how-to-do-it/22956 "2025-08-07T10:26:13Z")
In the documentation it is written: Trial Schedulers can … alter hyperparameters of a running trial. However it seems there is no scheduler class listed that does that. Am I missing something, do I understand someth… --- ## [How to log current Epsilon value on Ray Tune?](https://discuss.ray.io/t/how-to-log-current-epsilon-value-on-ray-tune/22797)
**Author:** [@armmer016](https://discuss.ray.io/u/armmer016)\ **Replies:** 2\ **Last updated:** [August 1, 2025, 7:56pm UTC](https://discuss.ray.io/t/how-to-log-current-epsilon-value-on-ray-tune/22797 "2025-08-01T19:56:44Z")
1. Severity of the issue: (select one) None: I’m just curious or want clarification. Low: Annoying but doesn’t hinder my work. Medium: Significantly affects my productivity but can find a workaround. High: Comple… --- ## [Use scheduler and search\_alg with different metrics](https://discuss.ray.io/t/use-scheduler-and-search-alg-with-different-metrics/22842)
**Author:** [@KB-Q](https://discuss.ray.io/u/KB-Q)\ **Replies:** 0\ **Last updated:** [July 14, 2025, 9:15am UTC](https://discuss.ray.io/t/use-scheduler-and-search-alg-with-different-metrics/22842 "2025-07-14T09:15:25Z")
1. Severity of the issue: (select one) \[\*\] High: Completely blocks me. 2. Environment: Ray version: 2.47.1 Python version: 3.11.11 OS: Linux Cloud/Infrastructure: Databricks jobs Other libs/tools (if relevant): optun… --- ## [Can SparkXGBoost tune with Ray tune?](https://discuss.ray.io/t/can-sparkxgboost-tune-with-ray-tune/22840)
**Author:** [@sornwit\_chotwatanapo](https://discuss.ray.io/u/sornwit_chotwatanapo)\ **Replies:** 0\ **Last updated:** [July 14, 2025, 7:24am UTC](https://discuss.ray.io/t/can-sparkxgboost-tune-with-ray-tune/22840 "2025-07-14T07:24:33Z")
I use Spark XGBoost in Databricks but I try to combine with ray tune but it doesn’t work Here is my code def train\_fn(config): train\_df = spark.table("Train\_df\_ray") val\_df = spark.table("Val\_df\_ray") fea… --- ## [How to use ray, hyperopt with OpenFOAM?](https://discuss.ray.io/t/how-to-use-ray-hyperopt-with-openfoam/22787)
**Author:** [@quarkz](https://discuss.ray.io/u/quarkz)\ **Replies:** 4\ **Last updated:** [July 10, 2025, 1:07am UTC](https://discuss.ray.io/t/how-to-use-ray-hyperopt-with-openfoam/22787 "2025-07-10T01:07:33Z")
1. Severity of the issue: (select one) Medium: Significantly affects my productivity but can find a workaround. 2. Environment: Ray version: 2.47 Python version: 3.10, 3.12 OS: win11 wsl Cloud/Infrastructure: na Ot… --- ## [Hyperopt with Ray Tune vs using Hyperopt directly](https://discuss.ray.io/t/hyperopt-with-ray-tune-vs-using-hyperopt-directly/922)
**Author:** [@Roger](https://discuss.ray.io/u/Roger)\ **Replies:** 3\ **Last updated:** [July 9, 2025, 3:50pm UTC](https://discuss.ray.io/t/hyperopt-with-ray-tune-vs-using-hyperopt-directly/922 "2025-07-09T15:50:42Z")
Hi, I want to use Hyperopt within Ray in order to parallelize the optimization and use all my computer resources. However, I found a difference in the behavior when running Hyperopt with Ray and Hyperopt library alone. … --- ## [Ray + Optuna + Grid Search + Pruner](https://discuss.ray.io/t/ray-optuna-grid-search-pruner/22792)
**Author:** [@Daraan](https://discuss.ray.io/u/Daraan)\ **Replies:** 1\ **Last updated:** [July 3, 2025, 8:37pm UTC](https://discuss.ray.io/t/ray-optuna-grid-search-pruner/22792 "2025-07-03T20:37:32Z")
I’ve written a lightweight custom Optuna searcher that allows to use ray.tune.grid\_search and pruners from optuna with it to early terminate trials like optuna would do. You can find it in my ray\_utilites project on Git… --- ## [Distributed training and distributed distributed training](https://discuss.ray.io/t/distributed-training-and-distributed-distributed-training/22773)
**Author:** [@habib-source](https://discuss.ray.io/u/habib-source)\ **Replies:** 0\ **Last updated:** [June 30, 2025, 8:15pm UTC](https://discuss.ray.io/t/distributed-training-and-distributed-distributed-training/22773 "2025-06-30T20:15:33Z")
1. Severity of the issue: (select one) None: I’m just curious or want clarification. 2. Environment: Ray version: 2.47.1 Python version: 3.12.8 OS: Arch Cloud/Infrastructure: Local In this tutorial: net = nn.Data… --- ## [Persisting checkpoints in Databricks](https://discuss.ray.io/t/persisting-checkpoints-in-databricks/22446)
**Author:** [@ishaan-mehta](https://discuss.ray.io/u/ishaan-mehta)\ **Replies:** 6\ **Last updated:** [June 4, 2025, 3:40pm UTC](https://discuss.ray.io/t/persisting-checkpoints-in-databricks/22446 "2025-06-04T15:40:05Z")
1. Severity of the issue: (select one) None: I’m just curious or want clarification. Low: Annoying but doesn’t hinder my work. Medium: Significantly affects my productivity but can find a workaround. High: Comple… --- ## [Continue training of finished trials (Tune, RLLIB, PPO)](https://discuss.ray.io/t/continue-training-of-finished-trials-tune-rllib-ppo/22505)
**Author:** [@davidhozic](https://discuss.ray.io/u/davidhozic)\ **Replies:** 3\ **Last updated:** [May 31, 2025, 9:38am UTC](https://discuss.ray.io/t/continue-training-of-finished-trials-tune-rllib-ppo/22505 "2025-05-31T09:38:16Z")
1. Severity of the issue: (select one) None: I’m just curious or want clarification. Low: Annoying but doesn’t hinder my work. Medium: Significantly affects my productivity but can find a workaround. High: Comple… --- ## [Building Ray from Source - update AxSearch](https://discuss.ray.io/t/building-ray-from-source-update-axsearch/21978)
**Author:** [@VMLC-PV](https://discuss.ray.io/u/VMLC-PV)\ **Replies:** 6\ **Last updated:** [May 28, 2025, 4:03am UTC](https://discuss.ray.io/t/building-ray-from-source-update-axsearch/21978 "2025-05-28T04:03:27Z")
How severe does this issue affect your experience of using Ray? Low: It annoys or frustrates me for a moment. I was trying to update ax\_search.py to make it work with newer versions of ax and I wanted to run the test … --- ## [Ray Tune gets stuck for infinity](https://discuss.ray.io/t/ray-tune-gets-stuck-for-infinity/22434)
**Author:** [@uiag](https://discuss.ray.io/u/uiag)\ **Replies:** 7\ **Last updated:** [May 5, 2025, 10:15pm UTC](https://discuss.ray.io/t/ray-tune-gets-stuck-for-infinity/22434 "2025-05-05T22:15:16Z")
1. Severity of the issue: (select one) \[ x\] High: Completely blocks me. 2. Environment: Ray version: 2.37.0 Python version: 3.12 OS: Windows Other libs/tools (if relevant): Ligthgbm 3. What happened vs. what you exp… --- ## [Add additional trials to experiment](https://discuss.ray.io/t/add-additional-trials-to-experiment/4338)
**Author:** [@MYousefi](https://discuss.ray.io/u/MYousefi)\ **Replies:** 5\ **Last updated:** [May 3, 2025, 9:07pm UTC](https://discuss.ray.io/t/add-additional-trials-to-experiment/4338 "2025-05-03T21:07:27Z")
This is my first project using Ray 1.6.0 and ray.tune Description: I am performing a network architecture search with ray tune. The call to run is given the parameters listed below. search\_alg = HyperOptSearch() if p… --- ## [Adding a node to a cluster while running tune experiments](https://discuss.ray.io/t/adding-a-node-to-a-cluster-while-running-tune-experiments/22269)
**Author:** [@matters](https://discuss.ray.io/u/matters)\ **Replies:** 0\ **Last updated:** [April 10, 2025, 2:40am UTC](https://discuss.ray.io/t/adding-a-node-to-a-cluster-while-running-tune-experiments/22269 "2025-04-10T02:40:49Z")
1. Severity of the issue: (select one) None: I’m just curious or want clarification. Low: Annoying but doesn’t hinder my work. Medium: Significantly affects my productivity but can find a workaround. High: Comple… --- ## [Relationship between Ray Workers and trials and CPUs](https://discuss.ray.io/t/relationship-between-ray-workers-and-trials-and-cpus/22177)
**Author:** [@ishaan-mehta](https://discuss.ray.io/u/ishaan-mehta)\ **Replies:** 7\ **Last updated:** [April 7, 2025, 10:01pm UTC](https://discuss.ray.io/t/relationship-between-ray-workers-and-trials-and-cpus/22177 "2025-04-07T22:01:24Z")
1. Severity of the issue: (select one) None: I’m just curious or want clarification. Low: Annoying but doesn’t hinder my work. Medium: Significantly affects my productivity but can find a workaround. High: Comple… [Next page](https://discuss.ray.io/c/ray-tune/5.md?page=1)