#retrodigitization — Public Fediverse posts
Live and recent posts from across the Fediverse tagged #retrodigitization, aggregated by home.social.
-
It just occurred to me that the ELECTRA training objective is quite similar to OCR post-correction tasks. Do you agree? And if this is so, what does follow from it? Should we be training ELECTRA models on historical documents? (I mean rather than BERT models, and geared specifically towards this post-correction task.)
-
It just occurred to me that the ELECTRA training objective is quite similar to OCR post-correction tasks. Do you agree? And if this is so, what does follow from it? Should we be training ELECTRA models on historical documents? (I mean rather than BERT models, and geared specifically towards this post-correction task.)
-
It just occurred to me that the ELECTRA training objective is quite similar to OCR post-correction tasks. Do you agree? And if this is so, what does follow from it? Should we be training ELECTRA models on historical documents? (I mean rather than BERT models, and geared specifically towards this post-correction task.)
-
It just occurred to me that the ELECTRA training objective is quite similar to OCR post-correction tasks. Do you agree? And if this is so, what does follow from it? Should we be training ELECTRA models on historical documents? (I mean rather than BERT models, and geared specifically towards this post-correction task.)
-
It just occurred to me that the ELECTRA training objective is quite similar to OCR post-correction tasks. Do you agree? And if this is so, what does follow from it? Should we be training ELECTRA models on historical documents? (I mean rather than BERT models, and geared specifically towards this post-correction task.)