Fig. 2
From: Hierarchical shared transfer learning for biomedical named entity recognition
![Fig. 2](http://media.springernature.com/full/springer-static/image/art%3A10.1186%2Fs12859-021-04551-4/MediaObjects/12859_2021_4551_Fig2_HTML.png)
Model segmentation schematics for hierarchical shared transfer learning. We split XLNet-CRF. The underlying layers of the token embedding and XLNet models are share, with the upper layers and CRFs of the XLNet as specific tasks