from transformers import RobertaTokenizerFast, RobertaForTokenClassification
from attacut import tokenize
import torch

tokenizer = RobertaTokenizerFast.from_pretrained("lst-nectec/HoogBERTa-SENTENCE-lst20")
model = RobertaForTokenClassification.from_pretrained("lst-nectec/HoogBERTa-SENTENCE-lst20")

To do Sentence Boundary Classification, use the following commands

python

from transformers import pipeline

nlp = pipeline('token-classification', model=model, tokenizer=tokenizer, aggregation_strategy="none")

sentence = "วันที่ 12 มีนาคมนี้ ฉันจะไปเที่ยววัดพระแก้ว ที่กรุงเทพ"
all_sent = []
sentences = sentence.split(" ")
for sent in sentences:
    all_sent.append(" ".join(tokenize(sent)).replace("_","[!und:]"))

sentence = " _ ".join(all_sent)

print(nlp(sentence))

For batch processing,

python

from transformers import pipeline

nlp = pipeline('token-classification', model=model, tokenizer=tokenizer, aggregation_strategy="none")

sentenceL = ["วันที่ 12 มีนาคมนี้","ฉันจะไปเที่ยววัดพระแก้ว ที่กรุงเทพ"]
inputList = []
for sentX in sentenceL:
  sentences = sentX.split(" ")
  all_sent = []
  for sent in sentences:
      all_sent.append(" ".join(tokenize(sent)).replace("_","[!und:]"))

  sentence = " _ ".join(all_sent)
  inputList.append(sentence)

print(nlp(inputList))

Huggingface Models

HoogBERTaEncoder

HoogBERTa: Feature Extraction and Mask Language Modeling

HoogBERTaMuliTaskTagger:

HoogBERTa-NER-lst20: Named-entity recognition (NER) based on LST20
HoogBERTa-POS-lst20: Part-of-speech tagging (POS) based on LST20
HoogBERTa-SENTENCE-lst20: Clause Boundary Classification based on LST20

Citation

Please cite as:

bibtex

@inproceedings{porkaew2021hoogberta,
  title = {HoogBERTa: Multi-task Sequence Labeling using Thai Pretrained Language Representation},
  author = {Peerachet Porkaew, Prachya Boonkwan and Thepchai Supnithi},
  booktitle = {The Joint International Symposium on Artificial Intelligence and Natural Language Processing (iSAI-NLP 2021)},
  year = {2021},
  address={Online}
}

Download full-text PDF
Check out the code on Github