File size: 356 Bytes
57bdca5
 
 
 
1
2
3
4
With our sliding window approach, however, there is overlap in
the tokens we pass to the model at each iteration. We don't want the log-likelihood for the tokens we're just treating
as context to be included in our loss, so we can set these targets to -100 so that they are ignored. The following
is an example of how we could do this with a stride of 512.