Add description

app.py

@@ -102,7 +102,16 @@ def predict(img, text):
 
     return prediction_vqa, prediction_aux
 
+description = """
+Less Is More: Linear Layers on CLIP Features as Powerful VizWiz Model
+
+Our approach focuses on visual question answering for visual impaired people. We fine-tuned our approach on the <a href='https://vizwiz.org/tasks-and-datasets/vqa/' >CVPR Grand Challenge VizWiz 2022</a> data set. 
+
+You may click on of the examples or upload your own image and question. The Gradio app shows the current answer for your question and an answer category.
+"""
+
 gr.Interface(fn=predict, 
+             description=description,
              inputs=[gr.Image(label='Image'), gr.Textbox(label='Question')],
              outputs=[gr.outputs.Label(label='Answer', num_top_classes=5), gr.outputs.Label(label='Answer Category', num_top_classes=7)],
              examples=[['examples/Augustiner.jpg', 'What is this?'],['examples/VizWiz_test_00006968.jpg', 'Can you tell me the color of the dog?'], ['examples/VizWiz_test_00005604.jpg', 'What drink is this?'], ['examples/VizWiz_test_00006246.jpg', 'Can you please tell me what kind of tea this is?'], ['examples/VizWiz_train_00004056.jpg', 'Is that a beer or a coke?'], ['examples/VizWiz_train_00017146.jpg', 'Can you tell me what\'s on this envelope please?'], ['examples/VizWiz_val_00003077.jpg', 'What is this?']]