This post uses pytorch-lightning v0.6.0 (PyTorch v1.3.1)and optuna v1.1.0.
Optuna is a hyperparameter optimization framework applicable to machine learning frameworks and black-box optimization solvers. PyTorch Lightning provides a lightweight PyTorch wrapper for better scaling with less code. Combining the two of them allows for automatic tuning of hyperparameters to find the best performing models.
Optuna is a black-box optimizer, which means it needs an objectivefunction, which returns a numerical value to evaluate the performance of the hyperparameters, and decide where to sample in upcoming trials.
In our example, we will be doing this for identifying MNIST characters from the Optuna GitHub examples folder. In this case, the objective function starts like this:
Notice that the objective function is passed an Optuna specific argument of trial. This object is passed to the objective function to be used to specify which hyperparameters should be tuned. This returns the accuracy of the model as return logger.metrics[-1][“accuracy”], which is used by Optuna as feedback on the performance of the trial.
Similar to how PyTorch uses Eager execution, Optuna allows you to define the kinds and ranges of hyperparameters you want to tune directly within your code using the trial object. This saves the effort of learning specialized syntax for hyperparameters, and also means you can use normal Python code to loop through or define your hyperparameters.
Optuna supports a variety of hyperparameter settings, which can be used to optimize floats, integers, or discrete categorical values. Numerical values can be suggested from a logarithmic continuum as well. In our MNIST example, we optimize the hyperparameters here:
The number of layers to be tuned is given from trial.suggest_int(“n_layers”, 1, 3), which gives an integer value from one to three, which will be labelled in Optuna as n_layers.
The dropout percentage is defined by trial.suggest_uniform(“dropout”, 0.2, 0.5), which gives a float value between 0.2 and 0.5.
For hyperparameters which should vary by orders of magnitude, such as learning rates, use something like trial.suggest_loguniform('learning_rate', 1e-5, 1000), which will vary the values from .00001 to 0.1.
Categorical selection from a list is possible with trial.suggest_categorical(‘optimizer’, [‘SGD’, ‘Adam’]).
The default sampler in Optuna Tree-structured Parzen Estimater (TPE), which is a form of Bayesian Optimization. Optuna uses TPE to search more efficiently than a random search, by choosing points closer to previous good results.
To run the trials, create a study object, which sets the direction of optimization ("maximize" or "minimize"), along with other settings. Then, the study object run with optimize(objective, n_trials=100, timeout=600), to do one hundred trials, with a timeout of ten minutes.
Each trial is chosen after evaluating all the trials that have been previously done, using a sampler to make a smart guess where the best values hyperparameters can be found. Optuna provides Tree-structured Parzen Estimator (TPE) samplers, which is a kind of bayesian optimization, as the default sampler.
The best values from the trials can be accessed through study.best_trial, and other methods of viewing the trials, such as formatting in a dataframe, are available.
Pruning trials is a form of early-stopping which terminates unpromising trials, so that computing time can be used for trials that show more potential. In order to do pruning, it’s necessary to open up the black-box of the Objective function some more to provide intermittent feedback on how the trial is going to Optuna, so it can compare the progress with the progress of other trials, and decide whether to stop the trial early, and provide a method to receive a method from Optuna when the trial should be terminated, and also allow the trial in session to terminate cleanly after recording the results. Fortunately, Optuna provides an integration for PyTorch Lightning (PyTorchLightingPruningCallBack) pruning that provides all of these functions.
After importing the PyTorchLighntingPruningCallback, passing it as a early_stop_callback to the trainer allows Lightning to do the pruning. The monitor argument of the PyTorchLighntingPruningCallback function references the PyTorch Lightning LightningModule dictionary and could be used for other entries, such as val_loss or val_acc.
For those interested, Optuna has many other features, including a visualizations, alternative samplers, optimizers, and pruning algorithms, as well as the ability to create user-defined versions as well. If you have more computing resources available, Optuna provides an easy interface for parallel trials to increase tuning speed.
Give Optuna a try!
The Ultimate Pytorch Research Framework
