Pretraining with GVP

[empyriumz]

Hi there,

I recently came across your work and was intrigued by the approach and findings. It's great that you've open-sourced the code.

I noticed using GVP as the backbone model for pretraining in Table 10 of the manuscript. Could you kindly share the configuration files used for the pretraining? I'm also curious to know whether this pretraining process with GVP uses the same 2-stage methodology as for the GearNet.

Thank you in advance for your help.