activebus/BERT-DK_rest

BERT-DK_rest is trained from 1G (19 types) restaurants from Yelp . The original model is from BERT-base-uncased trained from Wikipedia+BookCorpus . The post-trained weights are as simple as importing AutoModel, AutoTokenizer and Tokenizer . Check our NAACL paper below: "BERT Post-Training for Review Reading Comprehension and Aspect-based Sentiment Analysis", "ReviewBERT (post-)trained from review corpus to understand sentiment, options and various e-commence aspects", "Yelight", "Bert" and "Bert" are all post-training models that are used to help users understand,