BaseSb3AlgorithmSettings

Full path: schola.scripts.sb3.train.settings.BaseSb3AlgorithmSettings

Shared rollout and optimizer settings for on-policy SB3 algorithms.

BaseSb3AlgorithmSettings(learning_rate = 0.0003, n_steps = 2048,
                         batch_size = 64)

Parameters

• learning_rate (Annotated)

• n_steps (Annotated)

• batch_size (Annotated)

Methods

init

__init__(learning_rate = 0.0003, n_steps = 2048, batch_size = 64)

Parameters

• learning_rate (Annotated)

• n_steps (Annotated)

• batch_size (Annotated)

Attributes

batch_size

batch_size

Minibatch size for each update. This is the number of timesteps used in each batch for training the policy. Must be a divisor of n_steps.

---

learning_rate

learning_rate

Learning rate for the optimizer.

---

n_steps

n_steps

Number of steps to run for each environment per update. This is the number of timesteps collected before updating the policy.