"Environment Steps" and "Total Timesteps" - Githubissues

carlosferrazza / humanoid-bench

Other

308 stars 31 forks source link

"Environment Steps" and "Total Timesteps" #9

Open Ericongic opened 1 month ago

Ericongic commented 1 month ago

Dear Author,

I have a question regarding the relationship between the x-axis labels in Figures 5 and 6 of your paper and the output from the PPO algorithm in the repository.

What is the relationship between the x-axis label "environment steps" in Figures 5 and 6 of the paper and the "total_timesteps" output shown in the terminal when running the PPO algorithm from this repository?

How are "environment steps" and "total_timesteps" defined and calculated specifically in this context?

Thank you for your help.

carlosferrazza commented 3 weeks ago

Hi! They are the same thing. They correspond to the total number of environment transitions across all parallel environments.