Evaluating policy trained on simulation data.

[g0kul6]

I have trained the diffusion policy using the Libero dataset, and now I want to evaluate it. Is there any evaluation pipeline that you guys will recommend? (I have seen the one in the OpenVLA code base, but that is for the OpenVLA model. Since this involves multiple horizons for action, prediction, and observation, I was wondering if there is any standardized way to do this.)

[yfpeng1234]

Hi @g0kul6 , could you please share how to modify this codebase to train on libero? Thanks