Question about input
reluuu opened this issue · comments
Hi Thomas,
Thanks for the great work. I have a question about the batches. I noticed that there are two variables in inputs dict, in_batch and out_batch, I was wondering what is the difference between them? in_batch is the num of balls that we input KPConv, then what does out_batch represent? Why do they have different values?
self.inputs['in_batches'] = flat_inputs[ind]
ind += 1
self.inputs['out_batches'] = flat_inputs[ind]
Could you please help me solve this problem?
Best regards!
in_batch is a list containing the number of point in each ball in the current batch. The length of the list is the number of balls, and the elements are the number of points.
out_batch is the same but contains the number of points in the balls after the last subsampling layer. We need this value in the case of ModelNet classification to perform the global pooling and other debugging stuff.