Upload IntentAnalyzer v1.0 - Multi-Label Communication Intent Detection Model

README.md

@@ -55,7 +55,7 @@ IntentAnalyzer is a state-of-the-art multi-label text classification model desig
 The model detects 6 different intent categories (multi-label):
 
 1. **🧌 Trolling** - Deliberately provocative or disruptive communication
-2. **🚫 Dismissive** - Shutting down conversation or avoiding engagement  
+2. **🚫 Dismissive** - Shutting down conversation or avoiding engagement
 3. **🎭 Manipulative** - Using emotional coercion, guilt, or pressure tactics
 4. **🌋 Emotionally Reactive** - Overwhelmed by emotion, not thinking clearly
 5. **✅ Constructive** - Good faith engagement and dialogue
@@ -69,7 +69,7 @@ The model detects 6 different intent categories (multi-label):
 
 ### Per-Category Performance
 - **Trolling**: F1=0.943 (P=0.976, R=0.911)
-- **Dismissive**: F1=0.850 (P=0.964, R=0.761)  
+- **Dismissive**: F1=0.850 (P=0.964, R=0.761)
 - **Manipulative**: F1=0.907 (P=0.867, R=0.951)
 - **Emotionally Reactive**: F1=0.939 (P=0.931, R=0.947)
 - **Constructive**: F1=0.989 (P=0.978, R=1.000)
@@ -89,7 +89,7 @@ class MultiLabelIntentClassifier(nn.Module):
         self.bert = AutoModel.from_pretrained(model_name)
         self.dropout = nn.Dropout(0.3)
         self.classifier = nn.Linear(self.bert.config.hidden_size, num_labels)
-        
+
     def forward(self, input_ids, attention_mask):
         outputs = self.bert(input_ids=input_ids, attention_mask=attention_mask)
         pooled_output = outputs.last_hidden_state[:, 0]
@@ -110,18 +110,18 @@ intent_categories = ['trolling', 'dismissive', 'manipulative', 'emotionally_reac
 
 def predict_intent(text, threshold=0.5):
     inputs = tokenizer(text, return_tensors="pt", truncation=True, padding=True, max_length=128)
-    
+
     with torch.no_grad():
         outputs = model(inputs['input_ids'], inputs['attention_mask'])
         probabilities = torch.sigmoid(outputs).numpy()[0]
-    
+
     # Return predictions above threshold
     predictions = {}
     for i, category in enumerate(intent_categories):
         prob = probabilities[i]
         if prob > threshold:
             predictions[category] = prob
-            
+
     return predictions
 
 # Example usage
@@ -155,7 +155,7 @@ The model was trained on a carefully curated dataset of 1,226 examples with:
 - **Relationship Counseling**: Understand communication patterns
 - **Content Moderation**: Flag problematic intent patterns
 
-### Research Applications  
+### Research Applications
 - **Psychology**: Study communication patterns and intentions
 - **Linguistics**: Analyze pragmatic aspects of language
 - **Social Sciences**: Understanding online discourse patterns

modeling_intent.py

@@ -9,7 +9,7 @@ class MultiLabelIntentClassifier(nn.Module):
         self.bert = AutoModel.from_pretrained(model_name)
         self.dropout = nn.Dropout(0.3)
         self.classifier = nn.Linear(self.bert.config.hidden_size, num_labels)
-        
+
     def forward(self, input_ids, attention_mask):
         outputs = self.bert(input_ids=input_ids, attention_mask=attention_mask)
         pooled_output = outputs.last_hidden_state[:, 0]  # Use [CLS] token

pytorch_model.bin

@@ -1 +1,3 @@
-version https://git-lfs.github.com/spec/v1 oid sha256:2d06c315ccc3129b4860283640d34e976d89dea0b1d026118358e53854e43e32 size 265508834
+version https://git-lfs.github.com/spec/v1
+oid sha256:76ff6f53405352d5cdb384c2abb6e685d03cea834654311c19d79ec3984a7e66
+size 265508834