LossWeights
python
LossWeights(ppo: float, bc: float, entropy: float = 0.01) -> NonePPO and BC loss coefficients for one rollout's minibatch updates.
Values
bc
python
bc: floatentropy
python
entropy: float = 0.01The entropy bonus coefficient for this rollout, read from the schedule at the current step.
ppo
python
ppo: float