Evaluating on more benchmarks

[LaFeuilleMorte]

Hi, thanks for open-sourcing this awesome work. It looks like you evaluated your model only on libero and robocassa. Did you do some evaluation on other more challenging benchmarks like RoboTwin2.0 and GM-100