Dense layer tensor shape question


I’m am trying to implement an actor-critic algorithm (more specifically ddpg) and I was wondering if I feed the data the right way. I’ve seen previously that dense layers often expect B x C x F shapes where B is batch size, C is Channel size and F is feature size although in my scenario I’m having just B x F and it seems to work (!?) Is there something I’m missing? I’m providing 2 screenshots of the critic forward function and the output where I’m printing various shapes.