Hyperparameter Optimization

What is Hyperparameter Optimization?

Hyperparameters are variables that directly control the behaviors of training algorithms, and have a significant effect on the performance of the resulting machine learning models. Finding the hyperparameter values that yield the best performing models can be complicated. Manually adjusting hyperparameters over the course of many training trials can be slow and tedious. Luckily, you can automate and boost hyperparameter optimization (HPO) with ClearML's HyperParameterOptimizer class.

ClearML's Hyperparameter Optimization

ClearML provides the HyperParameterOptimizer class, which takes care of the entire optimization process for users with a simple interface.

ClearML's approach to hyperparameter optimization is scalable, easy to set up and to manage, and it makes it easy to compare results.

Workflow

The preceding diagram demonstrates the typical flow of hyperparameter optimization where the parameters of a base task are optimized:

1. Configure an Optimization Task with a base task whose parameters will be optimized, and a set of parameter values to test
2. Clone the base task. Each clone's parameter is overridden with a value from the optimization task
3. Enqueue each clone for execution by a ClearML Agent
4. The Optimization Task records and monitors the cloned tasks' configuration and execution details, and returns a summary of the optimization results in tabular and parallel coordinate formats, and in a scalar plot.

Supported Optimizers

The HyperParameterOptimizer class contains ClearML's hyperparameter optimization modules. Its modular design enables using different optimizers, including existing software frameworks, enabling simple, accurate, and fast hyperparameter optimization.

• Optuna - automation.optuna.OptimizerOptuna. Optuna is the default optimizer in ClearML. It makes use of different samplers such as grid search, random, bayesian, and evolutionary algorithms. For more information, see the Optuna documentation.
• BOHB - automation.hpbandster.OptimizerBOHB. BOHB performs robust and efficient hyperparameter optimization at scale by combining the speed of Hyperband searches with the guidance and guarantees of convergence of Bayesian Optimization. For more information about HpBandSter BOHB, see the HpBandSter documentation and a code example.
• Random uniform sampling of hyperparameters - automation.RandomSearch.
• Full grid sampling strategy of every hyperparameter combination - automation.GridSearch.
• Custom - automation.optimization.SearchStrategy - Use a custom class and inherit from the ClearML automation base strategy class

Defining a Hyperparameter Optimization Search Example

1. Import ClearML's automation modules:

   from clearml.automation import UniformParameterRange, UniformIntegerParameterRange
   from clearml.automation import HyperParameterOptimizer
   from clearml.automation.optuna import OptimizerOptuna

2. Initialize the Task, which will be stored in ClearML Server when the code runs. After the code runs at least once, it can be reproduced, and the parameters can be tuned:

   from clearml import Task
   
   task = Task.init(
       project_name='Hyper-Parameter Optimization',
       task_name='Automatic Hyper-Parameter Optimization',
       task_type=Task.TaskTypes.optimizer,
       reuse_last_task_id=False
   )

3. Define the optimization configuration and resources budget:

   optimizer = HyperParameterOptimizer(
         # specifying the task to be optimized, task must be in system already so it can be cloned
         base_task_id=TEMPLATE_TASK_ID,  
         # setting the hyperparameters to optimize
         hyper_parameters=[
             UniformIntegerParameterRange('number_of_epochs', min_value=2, max_value=12, step_size=2),
             UniformIntegerParameterRange('batch_size', min_value=2, max_value=16, step_size=2),
             UniformParameterRange('dropout', min_value=0, max_value=0.5, step_size=0.05),
             UniformParameterRange('base_lr', min_value=0.00025, max_value=0.01, step_size=0.00025),
             ],
         # setting the objective metric we want to maximize/minimize
         objective_metric_title='accuracy',
         objective_metric_series='total',
         objective_metric_sign='max',  
   
         # setting optimizer  
         optimizer_class=OptimizerOptuna,
     
         # configuring optimization parameters
         execution_queue='default',  
         max_number_of_concurrent_tasks=2,  
         optimization_time_limit=60., 
         compute_time_limit=120, 
         total_max_jobs=20,  
         min_iteration_per_job=15000,  
         max_iteration_per_job=150000,  
         )
   Locating Task ID

   To locate the base task's ID, go to the task's info panel in the WebApp. The ID appears in the task header.

Optimizer Execution Options

The HyperParameterOptimizer provides options to launch the optimization tasks locally or through a ClearML queue. Start a HyperParameterOptimizer instance using either HyperParameterOptimizer.start() or HyperParameterOptimizer.start_locally(). Both methods run the optimizer controller locally. start() launches the base task clones through a queue specified when instantiating the controller, while start_locally() runs the tasks locally.

Remote Execution

You can also launch the optimizer controller through a queue by using Task.execute_remotely() before starting the optimizer.

Tutorial

Check out the Hyperparameter Optimization tutorial for a step-by-step guide.

SDK Reference

For detailed information, see the complete HyperParameterOptimizer SDK reference page.

CLI

ClearML also provides clearml-param-search, a CLI utility for managing the hyperparameter optimization process. See ClearML Param Search for more information.

UI Application

Pro Plan Offering

The ClearML HPO App is available under the ClearML Pro plan

ClearML provides the Hyperparameter Optimization GUI application for launching and managing the hyperparameter optimization process.