OakEmu-RL v0.1.0

Appearance

PPOActionBundle ​

python
PPOActionBundle(actions: torch.Tensor, log_probs: torch.Tensor, entropy: torch.Tensor, values: torch.Tensor) -> None

Policy sampling outputs plus value estimates.

Source

Values ​

actions ​

python
actions: torch.Tensor

Sampled or supplied actions.

Source

entropy ​

python
entropy: torch.Tensor

Action-distribution entropy.

Source

log_probs ​

python
log_probs: torch.Tensor

Log probabilities of the selected actions.

Source

values ​

python
values: torch.Tensor

Associated value estimates.

Source