A simple plotting utility for algorithm performance comparison in reinforcement learning benchmarks. Plot the relative advantage of your algorithm against your baseline.
pip install wandb tqdm rladv
-
Run your algorithm and a baseline on a benchmark, logging the target performance metric to wandb.
-
Generate the plot:
import rladv
rladv.plot_advantage(project="my_project", comparison_variable="my_metric", cache=True, use_cached=True)Example Plot:
