search

VITA-Group / Open-L2O

Open-L2O: A Comprehensive and Reproducible Benchmark for Learning to Optimize Algorithms
MIT License
265 stars 45 forks source link

Hi there #1

Open microprediction opened 3 years ago

microprediction commented 3 years ago

I maintain the humpday package that compares global derivative-free optimizers. I'm curious to know how L2o approaches stack up against others, so just saying hi. I think it is a nice benchmark for more ambitious claims about RL in general. However, I'm sort of stuck on how to do this. What is a fair way to compare L20 that has learned Rastrigin against SHGO or whatever that has not?

lalalajiao commented 1 year ago

Did you run the ModelBase code successfully? Can we talk about it?

Lhongpei commented 1 year ago

Did you run the ModelBase code successfully? Can we talk about it?

Can you run it now? How did you solve error?