Ray Tune Examples

Tip

Check out the Tune User Guides To learn more about Tune’s features in depth.

Examples using Ray Tune with ML Frameworks

Ray Tune integrates with many popular machine learning frameworks. Here you find a few practical examples showing you how to tune your models. At the end of these guides you will often find links to even more examples.

How To Use Tune’s Scikit-Learn Adapters?

How To Use Tune With Keras & TF Models

How To Use Tune With PyTorch Models

How To Tune PyTorch Lightning Models

Model Selection & Serving With Ray Serve

Tuning RL Experiments With Ray Tune & Ray Serve

A Guide To Tuning XGBoost Parameters With Tune

A Guide To Tuning LightGBM Parameters With Tune

A Guide To Tuning Horovod Parameters With Tune

A Guide To Tuning Huggingface Transformers With Tune

End-to-end Example for Tuning a TensorFlow Model

End-to-end Example for Tuning a PyTorch Model with PBT

Tune Experiment Tracking Examples

Ray Tune integrates with some popular Experiment tracking and management tools, such as CometML, or Weights & Biases. If you’re interested in learning how to use Ray Tune with Tensorboard, you can find more information in our Guide to logging and outputs.

Using Aim with Ray Tune For Experiment Management

Using Comet with Ray Tune For Experiment Management

Tracking Your Experiment Process Weights & Biases

Using MLflow Tracking & AutoLogging with Tune

Tune Hyperparameter Optimization Framework Examples

Tune integrates with a wide variety of hyperparameter optimization frameworks and their respective search algorithms. Here you can find detailed examples on each of our integrations:

How To Use Tune With Ax

How To Use Tune With Dragonfly

How To Use Tune With HyperOpt

How To Use Tune With BayesOpt

How To Use Tune With BlendSearch and CFO

How To Use Tune With TuneBOHB

How To Use Tune With Nevergrad

How To Use Tune With Optuna

How To Use Tune With SigOpt

Other Examples

• tune_basic_example: Simple example for doing a basic random and grid search.

• Asynchronous HyperBand Example: Example of using a simple tuning function with AsyncHyperBandScheduler.

• HyperBand Function Example: Example of using a Trainable function with HyperBandScheduler. Also uses the AsyncHyperBandScheduler.

• Visualizing Population Based Training (PBT) Hyperparameter Optimization: Configuring and running (synchronous) PBT and understanding the underlying algorithm behavior with a simple example.

• PBT Function Example: Example of using the function API with a PopulationBasedTraining scheduler.

• PB2 Example: Example of using the Population-based Bandits (PB2) scheduler.

• Logging Example: Example of custom loggers and custom trial directory naming.

Tune Exercises

Learn how to use Tune in your browser with the following Colab-based exercises.

Exercise Description	Library	Colab Link
Basics of using Tune.	TF/Keras
Using Search algorithms and Trial Schedulers to optimize your model.	Pytorch
Using Population-Based Training (PBT).	Pytorch
Fine-tuning Huggingface Transformers with PBT.	Huggingface Transformers/Pytorch
Logging Tune Runs to Comet ML.	Comet

Tutorial source files can be found here.