Sharing decoder weights of VC-1

[godnpeter]

Hi, thank you guys for your great work!
I want to try out VC-1 on various tasks and am currently curious about how well it is able to reconstruct the current environments it is currently tested on.

I understand that VC-1 is mainly pre-trained on ego-centric video datasets, but still I think that seeing how well it is able to reconstruct a well masked image of the robotic simulators we are currently using is an interesting topic so I want to try this out.

However, it seems like that the weights for encoder of VC-1 is only available, so I hope you guys will be able to share the decoder weights as well!
Will this be possible?