serdarakyol commited on
Commit
7bf40a4
1 Parent(s): 4ce654d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +8 -1
README.md CHANGED
@@ -20,6 +20,8 @@ tokenizer = AutoTokenizer.from_pretrained("serdarakyol/interpress-turkish-news-c
20
 
21
  model = AutoModelForSequenceClassification.from_pretrained("serdarakyol/interpress-turkish-news-classification")
22
  ```
 
 
23
  ```sh
24
  # PREPROCESSING
25
  import re
@@ -91,7 +93,11 @@ def prediction(news):
91
  pred = np.argmax(logits,axis=1)[0]
92
  return pred
93
  ```
94
-
 
 
 
 
95
  ```sh
96
  labels = {
97
  0 : "Culture-Art",
@@ -107,6 +113,7 @@ labels = {
107
  }
108
  pred = prediction(news)
109
  print(labels[pred])
 
110
  ```
111
  Thanks to @yavuzkomecoglu for contributes
112
 
 
20
 
21
  model = AutoModelForSequenceClassification.from_pretrained("serdarakyol/interpress-turkish-news-classification")
22
  ```
23
+ ## NOTE: Please remember, for predict on BERT model, you don't actually need to preprocessing but the dataset was real world data. That why I needed to do some preprocessing. If you have normal news from any news web page, you can just copy the news and past. Then delete the first comment on ***prediction*** function. That's it.
24
+
25
  ```sh
26
  # PREPROCESSING
27
  import re
 
93
  pred = np.argmax(logits,axis=1)[0]
94
  return pred
95
  ```
96
+ ```sh
97
+ news = r"ABD'den Prens Selman'a yaptırım yok Beyaz Saray Sözcüsü Psaki, Muhammed bin Selman'a yaptırım uygulamamanın \"doğru karar\" olduğunu savundu. Psaki, \"Tarihimizde, Demokrat ve Cumhuriyetçi başkanların yönetimlerinde diplomatik ilişki içinde olduğumuz ülkelerin liderlerine yönelik yaptırım getirilmemiştir\" dedi."
98
+ ```
99
+ You can find the news in this [link](https://www.ntv.com.tr/dunya/abdden-prens-selmana-yaptirim-yok,YTeWNv0-oU6Glbhnpjs1JQ)
100
+ news data: 02/03/2021
101
  ```sh
102
  labels = {
103
  0 : "Culture-Art",
 
113
  }
114
  pred = prediction(news)
115
  print(labels[pred])
116
+ # > World
117
  ```
118
  Thanks to @yavuzkomecoglu for contributes
119