INNER CODE UNIT · Python

BERTLM

HendrikStrobelt/detecting-fake-text · backend/api.py:183

class BERTLM(AbstractLanguageChecker):
    def __init__(self, model_name_or_path="bert-base-cased"):
        super(BERTLM, self).__init__()
        self.device = torch.device(
            "cuda" if torch.cuda.is_available() else "cpu")
        self.tokenizer = BertTokenizer.from_pretrained(
            model_name_or_path,
            do_lower_case=False)
        self.model = BertForMaskedLM.from_pretrained(
            model_name_or_path)
        self.model.to(self.device)
        self.model.eval()
        # BERT-specific symbols
        self.mask_tok = self.tokenizer.convert_tokens_to_ids(["[MASK]"])[0]
        self.pad = self.tokenizer.convert_tokens_to_ids(["[PAD]"])[0]
        print("Loaded BERT model!")

    def check_probabilities(self, in_text, topk=40, max_context=20,

View source record →