Views
No views yet
koelectra-small-discriminator)1>>> from transformers import ElectraModel, ElectraTokenizer
2
3>>> model = ElectraModel.from_pretrained("monologg/koelectra-small-discriminator")
4>>> tokenizer = ElectraTokenizer.from_pretrained("monologg/koelectra-small-discriminator")1>>> from transformers import ElectraTokenizer
2>>> tokenizer = ElectraTokenizer.from_pretrained("monologg/koelectra-small-discriminator")
3>>> tokenizer.tokenize("[CLS] 한국어 ELECTRA를 공유합니다. [SEP]")
4['[CLS]', '한국어', 'E', '##L', '##EC', '##T', '##RA', '##를', '공유', '##합니다', '.', '[SEP]']
5>>> tokenizer.convert_tokens_to_ids(['[CLS]', '한국어', 'E', '##L', '##EC', '##T', '##RA', '##를', '공유', '##합니다', '.', '[SEP]'])
6[2, 18429, 41, 6240, 15229, 6204, 20894, 5689, 12622, 10690, 18, 3]1import torch
2from transformers import ElectraForPreTraining, ElectraTokenizer
3
4discriminator = ElectraForPreTraining.from_pretrained("monologg/koelectra-small-discriminator")
5tokenizer = ElectraTokenizer.from_pretrained("monologg/koelectra-small-discriminator")
6
7sentence = "나는 방금 밥을 먹었다."
8fake_sentence = "나는 내일 밥을 먹었다."
9
10fake_tokens = tokenizer.tokenize(fake_sentence)
11fake_inputs = tokenizer.encode(fake_sentence, return_tensors="pt")
12
13discriminator_outputs = discriminator(fake_inputs)
14predictions = torch.round((torch.sign(discriminator_outputs[0]) + 1) / 2)
15
16print(list(zip(fake_tokens, predictions.tolist()[1:-1])))