what does mean numbers in paranthesis of multi layer perceptron like (64,128,1024)

Question

Hello every one i'm reading pointnet papper but i cant understand the numbers of network architecture can you explain this to me. enter image description here

Answer 1

It means 3 fully connected layers, having neurons in each layer as 64, 128 and 1024 respectively. Between each layer, there is batch normalization as well as RELU activation (implemented in the paper).