Spaces:
Runtime error
Runtime error
Commit
·
b58669a
1
Parent(s):
3e39b83
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import re
|
3 |
+
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
|
4 |
+
from sheet2api import Sheet2APIClient
|
5 |
+
from datetime import datetime
|
6 |
+
from custom_functions import preprocessing_text
|
7 |
+
|
8 |
+
client = Sheet2APIClient(api_url='https://sheet2api.com/v1/hwp4AVQlOawy/summarizer')
|
9 |
+
|
10 |
+
def generate_summary(text):
|
11 |
+
|
12 |
+
client.create_row(row={'Timestamp': str(datetime.now())})
|
13 |
+
|
14 |
+
clean_text = preprocessing_text(text)
|
15 |
+
|
16 |
+
WHITESPACE_HANDLER = lambda k: re.sub('\s+', ' ', re.sub('\n+', ' ', k.strip()))
|
17 |
+
|
18 |
+
model_name = "csebuetnlp/mT5_multilingual_XLSum"
|
19 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name, use_fast=False) # use_fast was set to false
|
20 |
+
model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
|
21 |
+
|
22 |
+
input_ids = tokenizer(
|
23 |
+
[WHITESPACE_HANDLER(clean_text)],
|
24 |
+
return_tensors="pt",
|
25 |
+
padding="max_length",
|
26 |
+
truncation=True,
|
27 |
+
max_length=512
|
28 |
+
)["input_ids"]
|
29 |
+
|
30 |
+
output_ids = model.generate(
|
31 |
+
input_ids=input_ids,
|
32 |
+
max_length=84,
|
33 |
+
no_repeat_ngram_size=2,
|
34 |
+
num_beams=4
|
35 |
+
)[0]
|
36 |
+
|
37 |
+
summary = tokenizer.decode(
|
38 |
+
output_ids,
|
39 |
+
skip_special_tokens=True,
|
40 |
+
clean_up_tokenization_spaces=False
|
41 |
+
)
|
42 |
+
|
43 |
+
return summary
|
44 |
+
|
45 |
+
|
46 |
+
demo = gr.Interface(fn=generate_summary,
|
47 |
+
inputs=gr.Textbox(lines=10, placeholder="Insert the text here"),
|
48 |
+
outputs=gr.Textbox(lines=4)
|
49 |
+
)
|
50 |
+
|
51 |
+
demo.launch()
|