HumanCompatibleAI / overcooked_ai

A benchmark environment for fully cooperative human-AI performance.
https://arxiv.org/abs/1910.05789
MIT License
683 stars 144 forks source link

Cannot import env containing tomato task, order and ingredients #106

Closed KhoiDOO closed 1 year ago

KhoiDOO commented 1 year ago

Dear authors,

I'm trying to create an AgentEvaluator generated from the layout "asymmetric_advantages_tomato" however I get trouble in the problem shown below:

image

Could you pls show me if I have missed something, or how do I resolve this

Sincere!

micahcarroll commented 1 year ago

evaluate_human_model_pair uses planning-based (not learned) human models, but the planning code was not designed to be able to deal with anything other than 3-onion soups. Changing the planning code would be very challenging and we're not planning on doing so.

I'd recommend either using learned human models for this layout, or attempting to change the planning code (not recommended).