Reinforcement learning agent stops training unexpectedly?

3 visualizaciones (últimos 30 días)
lab
lab el 15 de Ag. de 2022
when I use DDPG agent to train a model, the agent stops training at 199th episode. However, the maximum episode number I set is 500. In addition, the average reward 143.9382 is far lower than the termination condition value 500. The final result shows "traning finished after all agents reached stop training criteria". I am confused what does it mean?

Respuestas (1)

Emmanouil Tzorakoleftherakis
Emmanouil Tzorakoleftherakis el 26 de En. de 2023
The picture also mentioned why training stopped in the last two rows. In this case, it seems you have a criterion that stops training when the average steps in an episode reach 500.
The max episode number kicks in if no other criterion is satisfied by the time training reaches the specified episode number.

Categorías

Más información sobre Training and Simulation en Help Center y File Exchange.

Productos


Versión

R2022a

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by