annotated-transformer icon indicating copy to clipboard operation
annotated-transformer copied to clipboard

Published 20 hours ago • verified publisher icon harvardnlp

nbatches vs batch_size

Open thomasahle opened this issue 1 year ago • 0 comments

In the forward method you currently say nbatches = query.size(0), but this appears to actually be the batch size, not the total number of batches the network is trained for.

I don't know if some people use this nomenclature differently, but I think renaming the variable to batch_size would be a bit more clear.

thomasahle avatar Nov 21 '22 21:11 thomasahle