Description
Training models by hand-tuning learning rate, layers and batch size burns nights per attempt. Optuna is an automatic hyperparameter optimization framework that efficiently finds the best settings.
Efficient sampling and pruning cut wasted trials, with distributed parallelism, visualization and support for major ML frameworks.
Auto-tuning:Smart search.
Pruning:Early stopping.
Distributed:Parallel trials.
Visualization:Analysis.
Efficient sampling and pruning cut wasted trials, with distributed parallelism, visualization and support for major ML frameworks.
Features
Auto-tuning:Smart search.
Pruning:Early stopping.
Distributed:Parallel trials.
Visualization:Analysis.
