Hi! How can I display the entropy of the policy network’s outputs, relative to the maximum possible entropy in Tensorboard?

@sven1977 is the right person to answer this question!

Hey @Lauritowal , you can specify a custom callback to calculate this property and store it for TB displaying. Like so: Check out the example script rllib/examples/custom_metrics_and_callbacks.py Override on_postprocess_trajectory and take a look at the postprocessed_batch arg therein. You should…

Entropy of policy network's output

RLlib

Ameer_Haj_Ali March 10, 2021, 9:35pm 2

CC @rliaw . Can you please help?

Adding custom metrics to policy

Topic		Replies	Views
Incredibly large policy entropy RLlib	3	387	November 13, 2021
Tensorboard Results RLlib	4	406	August 19, 2021
Use Policy_Trainer with TensorBoard RLlib	33	2497	November 13, 2021
Extract and display policy RLlib	3	533	July 26, 2021
How to directly use the custom_loss_model metric in tensorboard RLlib	2	409	June 11, 2021

Entropy of policy network's output

Related topics