I am running a APPO configuration, If i run with no “use_lstm”, I get these values after training for a while, but after adding lstm wrapper, I am not getting these values. Does it take more time to train with this wrapper? Does anyone know what is happening? New to Raycluster here.
INFO - Policy Loss: N/A
INFO - Value Function Loss: N/A
INFO - Entropy: N/A
INFO - Gradient Global Norm: N/A