The dependency parsing component was retrained using the [VnDT treebank v1.1](https://github.com/datquocnguyen/VnDT).
1.1
- NER model was retrained with additional VLSP 2018 data
1.0.1
Results published in our NAACL 2018 demo paper are associated with this version: * Word segmentation and POS tagging models were trained using the VLSP 2013 data * NER model was trained using the VLSP 2016 data * Dependency parsing model was trained using VnDT v1.0