HPO Builder (Optuna)

🔬 What is Optuna?

Optuna is an automatic hyperparameter optimization framework. It efficiently explores the search space using algorithms like TPE (Tree-structured Parzen Estimator) to find optimal hyperparameters faster than grid or random search.

📝 Study Configuration

Study Name

Number of Trials

Timesteps per Trial

1,000,000

Total Timesteps

~2-4 hrs

Est. Time (M4 Mac)

Search Space Size

🎲 Search Space

Toggle parameters to include in the search. Configure the range for each.

learning_rate

Min

Max

batch_size

Values (comma-separated)

n_steps

Values

gamma (discount)

gae_lambda

ent_coef (entropy)

Tip: Start with learning_rate, batch_size, and n_steps. These have the biggest impact on PPO performance. Add gamma and ent_coef if you have trial budget remaining.

💻 Generated Code

Copied to clipboard!