Question about the trained model

[Zijing5]

when I try to set cond_mode as action to add an extra embedding layer, I find the model you provided seems to lack the action condition embedding layer parameter. Will it be released sooner?Or I have missed some parts?