gabrielwithhappy commited on
Commit
e187645
โ€ข
1 Parent(s): c3bcc3d

fix null article exception

Browse files
Files changed (1) hide show
  1. app.py +16 -8
app.py CHANGED
@@ -17,12 +17,19 @@ def get_summary(input_text):
17
  summary_model = BartForConditionalGeneration.from_pretrained("ainize/kobart-news")
18
  input_ids = tokenizer.encode(input_text, return_tensors="pt")
19
  summary_text_ids = summary_model.generate(
20
- input_ids=input_ids,
21
- length_penalty=0.0,
22
- max_length=1000,
23
- min_length=56,
24
- num_beams=2,
25
  )
 
 
 
 
 
 
 
26
  return tokenizer.decode(summary_text_ids[0], skip_special_tokens=True)
27
 
28
 
@@ -60,7 +67,8 @@ class news_collector:
60
 
61
  for news_url in news_links:
62
  article = self.get_new_parser(news_url)
63
- self.examples_text.append([get_summary(article.text[:3000]), news_url])
 
64
 
65
 
66
 
@@ -97,8 +105,8 @@ with gr.Blocks(theme='pseudolab/huggingface-korea-theme') as demo:
97
 
98
  with gr.Tab("๋ฐ๋ชจ"):
99
  Link_TXT = gr.Textbox(label="๋‰ด์Šค ๋‚ด์šฉ", placeholder = "๋‰ด์Šค ๊ธฐ์‚ฌ ๋‚ด์šฉ์„ ์ž…๋ ฅํ•˜์„ธ์š”.")
100
- # gr.load("models/gabrielyang/finance_news_classifier-KR_v7",
101
- gr.load("models/Hyeonseo/ko-finance_news_classifier",
102
  inputs = Link_TXT)
103
  Link_URL = gr.Textbox(label="๋‰ด์Šค URL")
104
 
 
17
  summary_model = BartForConditionalGeneration.from_pretrained("ainize/kobart-news")
18
  input_ids = tokenizer.encode(input_text, return_tensors="pt")
19
  summary_text_ids = summary_model.generate(
20
+ length_penalty=0,
21
+ top_p=0.9,
22
+ max_length=128,
23
+ min_length=12,
24
+ num_beams=4,
25
  )
26
+ # "task_specific_params": {
27
+ # "summarization": {
28
+ # "length_penalty": 1.0,
29
+ # "max_length": 128,
30
+ # "min_length": 12,
31
+ # "num_beams": 4
32
+ # }
33
  return tokenizer.decode(summary_text_ids[0], skip_special_tokens=True)
34
 
35
 
 
67
 
68
  for news_url in news_links:
69
  article = self.get_new_parser(news_url)
70
+ if article.text:
71
+ self.examples_text.append([get_summary(article.text[:1500]), news_url])
72
 
73
 
74
 
 
105
 
106
  with gr.Tab("๋ฐ๋ชจ"):
107
  Link_TXT = gr.Textbox(label="๋‰ด์Šค ๋‚ด์šฉ", placeholder = "๋‰ด์Šค ๊ธฐ์‚ฌ ๋‚ด์šฉ์„ ์ž…๋ ฅํ•˜์„ธ์š”.")
108
+ gr.load("models/gabrielyang/finance_news_classifier-KR_v7",
109
+ # gr.load("models/Hyeonseo/ko-finance_news_classifier",
110
  inputs = Link_TXT)
111
  Link_URL = gr.Textbox(label="๋‰ด์Šค URL")
112