原书出处
Natural Language Processing in Action, Second Edition
Another popular approach in the Keras community is to retain the hidden layer from a previous batch of documents. This pretrained hidden layer embedding gives your language model information about the context of the new document—the text that came before it. However, this only makes sense if you’ve maintained the order of your documents within the batches and across the batches that you are training.
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-8.html
EPUB SHA-256:0e9a35e49ca446e322e5524e53f39fb6bee5b2b23d9cda10ef052aca770b0b36