You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
At this line, logits is copied from t_logits to logits_out, with offset = 0, size = n_outputs*n_vocab*sizeof(float). However, it seems that logits of output is not always contiguous at tensor memory, for example the picture at below:
reacted with thumbs up emoji reacted with thumbs down emoji reacted with laugh emoji reacted with hooray emoji reacted with confused emoji reacted with heart emoji reacted with rocket emoji reacted with eyes emoji
Uh oh!
There was an error while loading. Please reload this page.
-
At this line, logits is copied from t_logits to logits_out, with
offset = 0
,size = n_outputs*n_vocab*sizeof(float)
. However, it seems that logits of output is not always contiguous at tensor memory, for example the picture at below:So how does this case is handled?
Beta Was this translation helpful? Give feedback.
All reactions