RamiIbrahim commited on
Commit
ea775d8
1 Parent(s): f5fb067

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +34 -18
app.py CHANGED
@@ -46,37 +46,53 @@ formatted_examples = [
46
  # Create Gradio interface
47
  iface = gr.Interface(
48
  fn=predict_sentiment,
49
- inputs=gr.Textbox(lines=3, placeholder="Enter Tunisian Arabiz text here..."),
50
  outputs=[
51
- gr.Label(label="Predicted Sentiment"),
52
- gr.Label(label="Confidence Score"),
53
- gr.Textbox(label="Explanation")
54
  ],
55
  examples=formatted_examples,
56
- title="Tunisian Arabiz Sentiment Analysis",
57
  description="""
58
  <div style="text-align: center; margin-bottom: 20px;">
59
- <img src="cover.png" alt="Cover picture" style="width:300px;"/>
60
  </div>
61
- This model predicts the sentiment of Tunisian Arabiz text as either Positive or Negative.
62
 
63
- Tunisian Arabiz is a form of writing Arabic (specifically Tunisian dialect) using Latin characters and numbers.
64
 
65
- Example:
66
- - "3ajbetni" means "I liked it"
67
- - "7aja" means "thing"
68
- - "a3tini 9ahwa" means "give me a coffoe"
 
 
 
69
 
70
- Try the examples below or enter your own text!
 
71
  """,
72
  article="""
73
- <h3>About the Model</h3>
74
- <p>This sentiment analysis model was trained on a dataset combining TuniziDataset and the Tunisian Dialect Corpus.
75
  It uses TF-IDF vectorization for feature extraction and Logistic Regression for classification.</p>
76
- <p>It accepts both tunisian arabiz and actual arabic letters<p/>
77
- <h3>Limitations</h3>
 
 
 
 
 
 
 
 
 
78
  <p>The model may not perform well on very colloquial expressions or new slang terms not present in the training data.
79
- It's also important to note that sentiment can be nuanced and context-dependent, which may not always be captured by this model.</p>
 
 
 
80
  """
81
  )
82
 
 
46
  # Create Gradio interface
47
  iface = gr.Interface(
48
  fn=predict_sentiment,
49
+ inputs=gr.Textbox(lines=3, placeholder="أدخل النص هنا... / Enter your text here..."),
50
  outputs=[
51
+ gr.Label(label="Predicted Sentiment / التوقع"),
52
+ gr.Label(label="Confidence Score / نسبة الثقة"),
53
+ gr.Textbox(label="Explanation / الشرح")
54
  ],
55
  examples=formatted_examples,
56
+ title="Tunisian Arabiz & Arabic Sentiment Analysis / تحليل المشاعر للهجة التونسية",
57
  description="""
58
  <div style="text-align: center; margin-bottom: 20px;">
59
+ <img src="/cover.png" alt="Cover picture" style="width:300px;"/>
60
  </div>
61
+ <p>This model predicts the sentiment of Tunisian text as either Positive or Negative. It works with both Tunisian Arabiz and standard Arabic script.</p>
62
 
63
+ <p>هذا النموذج يتوقع مشاعر النص التونسي إما إيجابية أو سلبية. يعمل مع كل من الكتابة العربيزية التونسية والنص العربي القياسي.</p>
64
 
65
+ <h4>What is Tunisian Arabiz? / ما هي العربيزية التونسية؟</h4>
66
+ <p>Tunisian Arabiz is a way of writing the Tunisian dialect using Latin characters and numbers. For example:</p>
67
+ <ul>
68
+ <li>"3ajbetni" means "I liked it" / تعني "أعجبتني"</li>
69
+ <li>"7aja" means "thing" / تعني "حاجة"</li>
70
+ <li>"a3tini 9ahwa" means "give me a coffee" / تعني "أعطيني قهوة"</li>
71
+ </ul>
72
 
73
+ <p>Try the examples below or enter your own text!</p>
74
+ <p>!جرب الأمثلة أدناه أو أدخل نصك الخاص</p>
75
  """,
76
  article="""
77
+ <h3>About the Model / حول النموذج</h3>
78
+ <p>This sentiment analysis model was trained on a combined dataset from TuniziDataset and the Tunisian Dialect Corpus.
79
  It uses TF-IDF vectorization for feature extraction and Logistic Regression for classification.</p>
80
+
81
+ <p>تم تدريب نموذج تحليل المشاعر هذا على مجموعة بيانات مجمعة من TuniziDataset ومجموعة اللهجة التونسية.
82
+ يستخدم التجزئة TF-IDF لاستخراج الميزات والانحدار اللوجستي للتصنيف.</p>
83
+
84
+ <p>The model accepts both Tunisian Arabiz and standard Arabic script.</p>
85
+ <p>يقبل النموذج كلاً من النص العربيزي التونسي والنص العربي القياسي.</p>
86
+
87
+ <h3>Limitations / القيود</h3>
88
+ <p>Due to dataset limitations, neutral sentiment data was removed to achieve maximum performance.
89
+ This model is open-source, and contributions of additional datasets are welcome to improve its capabilities.</p>
90
+
91
  <p>The model may not perform well on very colloquial expressions or new slang terms not present in the training data.
92
+ Sentiment can be nuanced and context-dependent, which may not always be captured accurately by this model.</p>
93
+
94
+ <p>قد لا يؤدي النموذج بشكل جيد مع التعبيرات العامية جدًا أو المصطلحات الجديدة غير الموجودة في بيانات التدريب.
95
+ يمكن أن تكون المشاعر دقيقة ومعتمدة على السياق، وهو ما قد لا يتم التقاطه بدقة دائمًا بواسطة هذا النموذج.</p>
96
  """
97
  )
98