Regression Report: wandb

[['?we=costa-huang&wpn=trl&ceik=tracker_project_name&cen=log_with&metrics=env/reward_mean&metrics=objective/kl&metrics=objective/entropy', 'wandb?tag=calculator_few_shots_env&tag=pr-429&cl=calculator_env (various improvement)']]

Costa

Created on June 15|Last edited on June 15

Comment

﻿
﻿
env/reward_mean trl
env/reward_mean trl
100200300Steps00.10.20.30.40.50.6Episodic Return
objective/kl trl
objective/kl trl
100200300Steps-600-500-400-300-200-1000Episodic Return
objective/entropy trl
objective/entropy trl
100200300Steps100200300400500600700Episodic Return
env/reward_mean trl
env/reward_mean trl
10203040Time (minutes)00.20.40.60.81Episodic Return
objective/kl trl
objective/kl trl
10203040Time (minutes)-600-400-2000Episodic Return
objective/entropy trl
objective/entropy trl
10203040Time (minutes)200400600800Episodic Return
calculator_env (various improvement)10
﻿
﻿

Add a comment