File size: 832 Bytes
65f03a1 2241dbc 7767171 23cd87c 65f03a1 |
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 27 28 29 30 31 |
import os
import gradio as gr
import torch
from transformers.models.bert import BertTokenizer, BertForSequenceClassification
path='/home/user/app'
vocab='vocab.txt'
tokenizer = BertTokenizer.from_pretrained(os.path.join(path,vocab))
model = BertForSequenceClassification.from_pretrained(path)
model.eval()
def get_output(text):
output=[]
model_input = tokenizer(text, return_tensors="pt", padding=True)
model_output = model(**model_input, return_dict=False)
prediction = torch.argmax(model_output[0].cpu(), dim=-1)
prediction = [p.item() for p in prediction]
for i in range(len(prediction)):
if prediction[i]==0:
output.append("消极")
else:
output.append('积极')
return output
demo=gr.Interface(fn=get_output,inputs='text',outputs='text')
demo.launch() |