原书出处
Natural Language Processing in Action, Second Edition
The token sequence input into a transformer is always truncated and padded so that it has the required length, no matter how long or short the text being processed is. For example, the token sequence [What] [is] [BERT] [?] could be encoded as four embedding column vectors, creating an N×4 tensor or array. But this encoding would be too short for a sentence with five
原书定位详情
来源:a19c7933-4eda-4176-aebd-30fe7e1ad292
正文位置:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-9.html
EPUB SHA-256:0e9a35e49ca446e322e5524e53f39fb6bee5b2b23d9cda10ef052aca770b0b36