Spaces:
Sleeping
Sleeping
create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
|
3 |
+
from transformers import BertTokenizer, EncoderDecoderModel, EncoderDecoderConfig
|
4 |
+
model_ckpt = 'ardavey/bert2gpt-indosum'
|
5 |
+
tokenizer = BertTokenizer.from_pretrained(model_ckpt)
|
6 |
+
tokenizer.bos_token = tokenizer.cls_token
|
7 |
+
tokenizer.eos_token = tokenizer.sep_token
|
8 |
+
|
9 |
+
config = EncoderDecoderConfig.from_pretrained(model_ckpt)
|
10 |
+
config.early_stopping = True
|
11 |
+
|
12 |
+
model = EncoderDecoderModel.from_pretrained(model_ckpt, config=config)
|
13 |
+
|
14 |
+
text = st.text('Enter an article to summarize:')
|
15 |
+
|
16 |
+
if text:
|
17 |
+
input_ids = tokenizer.encode(custom_text, return_tensors='pt', padding=True, truncation=True, max_length=512)
|
18 |
+
summary_ids = model.generate(input_ids,
|
19 |
+
min_length=40,
|
20 |
+
max_length=200,
|
21 |
+
num_beams=10,
|
22 |
+
repetition_penalty=2.0,
|
23 |
+
length_penalty=1.0,
|
24 |
+
no_repeat_ngram_size=3,
|
25 |
+
use_cache=True,
|
26 |
+
do_sample = False,
|
27 |
+
top_k = 50,
|
28 |
+
)
|
29 |
+
|
30 |
+
summary_text = tokenizer.decode(summary_ids[0], skip_special_tokens=True)
|
31 |
+
# capitalize the first letter of the summary and after each period
|
32 |
+
def capitalize_sentences(text):
|
33 |
+
sentences = text.split('. ')
|
34 |
+
capitalized_sentences = [sentence[0].upper() + sentence[1:] if sentence else sentence for sentence in sentences]
|
35 |
+
return '. '.join(capitalized_sentences)
|
36 |
+
|
37 |
+
# correct any wrong terms using the replacement_dict
|
38 |
+
replacement_dict = {
|
39 |
+
"optiglain": "OptiGuard",
|
40 |
+
"telkom university": "Telkom University",
|
41 |
+
"menyerbut": "menyebut"
|
42 |
+
}
|
43 |
+
|
44 |
+
for wrong_term, correct_term in replacement_dict.items():
|
45 |
+
summary_text = summary_text.replace(wrong_term, correct_term)
|
46 |
+
|
47 |
+
summary_text = capitalize_sentences(summary_text)
|
48 |
+
st.info(summary_text)
|
49 |
+
|
50 |
+
|