What is the spatial conv?

What is the spatial conv?
I would like to create the model shown in the figure below. What are spatial and temporal conv? And what kind of pytorch should be structured?
Fig. 2. - Detailed nework architecture of the representation module.

Most likely it’s just a categorical meaning. You are probably using certain date which represents space and time. The author denotes the convolution as “temporal” meaning the kernel spawns in the temporal dim. Analogously to the spatial one.

PD: it’s just a 2d convolution.