blob: a4f9db552331469aa80cc79f9495f5f262fdfdfa (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
|
from transformers import BertModel, BertTokenizer
from transformers.models.bert import BertModel
import torch
from torch import nn
nn.BatchNorm2d()
if __name__ == '__main__':
model_name = 'bert-base-uncased'
tokenizer = BertTokenizer.from_pretrained(model_name)
model = BertModel.from_pretrained(model_name, output_hidden_states=True)
text = "After stealing money from the bank vault, the bank robber was seen " \
"fishing on the Mississippi river bank."
model.eval()
token_inputs = tokenizer(text, return_tensors='pt')
with torch.no_grad():
outputs = model(**token_inputs)
|
