search

wenwenyu / PICK-pytorch

Code for the paper "PICK: Processing Key Information Extraction from Documents using Improved Graph Learning-Convolutional Networks" (ICPR 2020)
https://arxiv.org/abs/2004.07464
MIT License
553 stars 191 forks source link

Do we need to train a seperate model for each dataset? #95

Open etrigger opened 3 years ago

etrigger commented 3 years ago

@wenwenyu Thanks for your greate work. For different datasets, do we have to train a seperate model for different datasets? Or we can mix all the different types of datasets and train a single model with a large number of entities.

AtulKumar4 commented 3 years ago

@etrigger If all the required entities are present in all the datasets and the datasets layout is similar then it's a good idea to train a single model.