• Retrain a new tokenization model on a much bigger dataset. F1 score =0.985
• Add training data and training code
• Better integration to spacy.io (removing redundant spaces between tokens after tokenization. Eg. Việt Nam , 12 / 22 / 2020 => Việt Nam, 12/22/2020]

• Tokenization
• POS tagging
• Accents removal
• Accents adding

Algorithm: Conditional Random Field

Vietnamese tokenizer f1_score = 0.985

Vietnamese pos tagging f1_score = 0.925

POS TAGS:

• A - Adjective
• C - Coordinating conjunction
• E - Preposition
• I - Interjection
• L - Determiner
• M - Numeral
• N - Common noun
• Nc - Noun Classifier
• Ny - Noun abbreviation
• Np - Proper noun
• Nu - Unit noun
• P - Pronoun
• R - Adverb
• S - Subordinating conjunction
• T - Auxiliary, modal words
• V - Verb
• X - Unknown
• F - Filtered out (punctuation)

At the command line with pip

$ pip install pyvi

Uninstall

$ pip uninstall pyvi

from pyvi import ViTokenizer, ViPosTagger
ViTokenizer.tokenize(u"Trường đại học bách khoa hà nội")
ViPosTagger.postagging(ViTokenizer.tokenize(u"Trường đại học Bách Khoa Hà Nội")
from pyvi import ViUtils
ViUtils.remove_accents(u"Trường đại học bách khoa hà nội")
from pyvi import ViUtils
ViUtils.add_accents(u'truong dai hoc bach khoa ha noi')