Problem with using LLM cache

[AaronChuh]

I really appreciate your work, but when I try to use the python scalingup/inference.py evaluation.num_episodes=50000 policy=scalingup evaluation=drawer to generate some trajetories, I got a WARNING that I didn't set the OPENAI_API_KEY. I've downloaded the cache file with this command wget https://www.cs.columbia.edu/~huy/scalingup/static/responses.zip and unzip the files into the project directory. However, I still got the same problem.
So could you tell me how to use the cache? Thanks!

[huy-ha]

I'm guessing this is the same issue as #22. Could you try checking out the first commit in the git history?