I've got point cloud data from meshes for skyrim npcs heads

I've converted them into sequences of integers for each head example.

I can easily convert that 1 dimensional sequence back and recreate the mesh.

I want to "use AI", to generate similar heads with variations.

I'm thinking about an easy solution, like why can't I just call a library in python that will take the dataframe with the 20 or so examples, and generate additional rows of synthetic cases?

Is there a really easy GAN that will just take a series of lists of integers and generate similar lists?

pic related, its what about 20 or so examples of heads look like point by point.

You can see that the position of each point is very similar for each head... how best to generate variations of heads from this data?