You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I want to evaluate my agent in webshop environment to calculate Task Score and Success Rate.
I successfully bought a product for an instruction and got a reward. However, I want to evaluate on 100 sample instructions and then calculate the average task score.
What I've tried:
I used env.reset() as it changes the instruction everytime as I have observed. However, I am not certain if it starts repeating the same instruction or not after a while. Is there any other function or method that I can call to change the instruction and knowing that it will not repeat the instruction that I've already executed on.
And while we're at it, is there any built-in way to calculate success rate?
By the way, wonderful work with the project and paper. Thanks in advance. @ysymyth
The text was updated successfully, but these errors were encountered:
What I want to do:
I want to evaluate my agent in webshop environment to calculate Task Score and Success Rate.
I successfully bought a product for an instruction and got a reward. However, I want to evaluate on 100 sample instructions and then calculate the average task score.
What I've tried:
I used
env.reset()
as it changes the instruction everytime as I have observed. However, I am not certain if it starts repeating the same instruction or not after a while. Is there any other function or method that I can call to change the instruction and knowing that it will not repeat the instruction that I've already executed on.And while we're at it, is there any built-in way to calculate success rate?
By the way, wonderful work with the project and paper. Thanks in advance. @ysymyth
The text was updated successfully, but these errors were encountered: