How to explain create_critic_learner architecture

[aamir3828 (Aamir Khan)]

Is there anyone that can explain about this architecture that how many convolution layers used and what is meant by conv1d there. I want to draw a figure for my research paper.