LibtorchRLPPOLossOutput Struct Reference

Actor, critic, and entropy terms produced by one PPO loss evaluation. More...

#include <LibtorchRLPPOLoss.h>

Public Attributes
torch::Tensor	actor_loss
	Clipped actor loss for the current mini-batch.
torch::Tensor	critic_loss
	Critic regression loss for the current mini-batch.
torch::Tensor	entropy
	Mean action-distribution entropy for the current mini-batch.

Detailed Description

Actor, critic, and entropy terms produced by one PPO loss evaluation.

Definition at line 21 of file LibtorchRLPPOLoss.h.

Member Data Documentation

actor_loss

torch::Tensor LibtorchRLPPOLossOutput::actor_loss

Clipped actor loss for the current mini-batch.

Definition at line 24 of file LibtorchRLPPOLoss.h.

Referenced by LibtorchRLPPOLoss::compute(), and LibtorchDRLControlTrainer::trainController().

critic_loss

torch::Tensor LibtorchRLPPOLossOutput::critic_loss

Critic regression loss for the current mini-batch.

Definition at line 26 of file LibtorchRLPPOLoss.h.

Referenced by LibtorchRLPPOLoss::compute().

entropy

torch::Tensor LibtorchRLPPOLossOutput::entropy

Mean action-distribution entropy for the current mini-batch.

Definition at line 28 of file LibtorchRLPPOLoss.h.

Referenced by LibtorchRLPPOLoss::compute().

---

The documentation for this struct was generated from the following file:

• stochastic_tools/include/libtorch/utils/LibtorchRLPPOLoss.h