lamda-bbo / offline-moo

Official implementation of ICML'24 paper "Offline Multi-Objective Optimization".
15 stars 4 forks source link

Running the baselines for synthetic functions #11

Open yannadani opened 2 months ago

yannadani commented 2 months ago

Thank you for the great paper and repo.

I am trying to run multi_head baselines on synthetic functions (for ex vlmop1). I haven't configured EvoXbench or MujoCO for now. But when I call bash scripts/multi_head.sh or the direct python command, it always tries to configure EvoXbench and throws an error that a module is missing. I was wondering if the data related to EvoXbench has to be downloaded and configured in order to run on synthetic functions as well. Or is there a simple way to install and configure such that I can run only on synthetic benchmark.

Any help would be great!

trxcc commented 2 months ago

Hi @yannadani

Thank you for your attention to our work! I have checked our implementation and reproduced your issue. An error would be raised since we directly import off_moo_bench.mo_nas in our implementation.

I have added some conditional checks in our code, which refers to #12. Now I can successfully run the algorithms (for example, run bash scripts/multi_head.sh directly).

If you have any questions, feel free to contact us!