Linear Evaluation LR Sweep Values

Hi, in the paper, you write "For each model, we sweep the learning rate value.". What are the swept learning rate values?

Thank you.