OakEmu-RL v0.1.0

Appearance

LossResult ​

python
LossResult(weighted_loss: torch.Tensor, td_errors: torch.Tensor, q_values: torch.Tensor, raw_loss: torch.Tensor) -> None

A computed DQN loss, ready for the caller to perform the optimizer step.

Source

Values ​

q_values ​

python
q_values: torch.Tensor

Source

raw_loss ​

python
raw_loss: torch.Tensor

Source

td_errors ​

python
td_errors: torch.Tensor

Source

weighted_loss ​

python
weighted_loss: torch.Tensor

Source