Update app.py

app.py

@@ -107,9 +107,6 @@ def predict_single_text(text):
     # Calculate the probabilities
     probabilities = torch.sigmoid(logits).squeeze()
 
-    # Define the threshold for prediction
-    threshold = 0.3
-
     # Get the predicted labels
     predicted_labels_ = (probabilities.cpu().numpy() > threshold).tolist()
 
@@ -162,12 +159,13 @@ def predict_single_text(text):
 
 # Create Gradio interface for single text
 iface2 = gr.Interface(fn=predict_single_text,
-                      inputs=gr.Textbox(lines=7, label="Paste or type text here"),
+                      inputs=[gr.Textbox(lines=7, label="Paste or type text here"),
+                              gr.Slider(minimum=0.0, maximum=1.0, value=0.3, step=0.01, label="Threshold value (default=0.3)")],
                       outputs=[gr.Label(label="Top Predictions", show_label=True),
                                gr.Plot(label="Likelihood of all labels", show_label=True)],
                       title="Single Text Prediction",
-                      article="**Note:** The quality of model predictions may depend on the quality of information provided."
-                      )
+                      description="**Threshold value:** The threshold value determines the minimum probability required for a label to be predicted. A higher threshold value will result in fewer labels being predicted, while a lower threshold value will result in more labels being predicted. The default threshold value is 0.3.",
+                      article="**Note:** The quality of model predictions may depend on the quality of the information provided.")
 
 
 # UPLOAD CSV
@@ -245,8 +243,8 @@ def predict_from_csv(file, column_name, progress=gr.Progress()):
         # Calculate the probabilities
         predictions = torch.sigmoid(logits).squeeze()
 
-        # Define the threshold for prediction
-        threshold = 0.3
+        # # Define the threshold for prediction
+        # threshold = 0.3
 
         # Get the predicted labels
         predicted_labels_ = (predictions.cpu().numpy() > threshold).tolist()
@@ -352,13 +350,17 @@ def predict_from_csv(file, column_name, progress=gr.Progress()):
 # Define the input component
 file_input = gr.File(label="Upload CSV or Excel file here", show_label=True, file_types=[".csv", ".xls", ".xlsx"])
 column_name_input = gr.Textbox(label="Enter the column name containing the text to be analyzed", show_label=True)
+threshold_input = gr.Slider(minimum=0.0, maximum=1.0, value=0.3, step=0.01, label="Threshold value (default=0.3)")
 
 # Create the Gradio interface
 iface3 = gr.Interface(fn=predict_from_csv,
-                      inputs=[file_input, column_name_input],
+                      inputs=[file_input, column_name_input, 
+                              gr.Slider(minimum=0.0, maximum=1.0, value=0.3, step=0.01, label="Threshold value (default=0.3)")],
                       outputs=gr.File(label='Download output CSV', show_label=True),
                       title="Multi-text Prediction",
-                      description='**NOTE:** Please enter the column name containing the text to be analyzed.')
+                      description='''**Threshold value:** The threshold value determines the minimum probability required 
+                      for a label to be predicted. A higher threshold value will result in fewer labels being predicted, 
+                      while a lower threshold value will result in more labels being predicted. The default threshold value is 0.3''')
 
 # Create a tabbed interface
 demo = gr.TabbedInterface(interface_list=[iface1, iface2, iface3],