Update app.py

app.py

@@ -33,14 +33,12 @@ processor = AutoProcessor.from_pretrained("Qwen/Qwen2-VL-2B-Instruct", trust_rem
 
 @spaces.GPU()
 def process_pdf_and_query(pdf_file, user_query):
-    # Convert the PDF to images
-    images = convert_from_path(pdf_file.name)  # pdf_file.name gives the file path
+    images = convert_from_path(pdf_file.name)  
     num_images = len(images)
 
-    # Indexing the PDF in RAG
     RAG.index(
         input_path=pdf_file.name,
-        index_name="image_index",  # index will be saved at index_root/index_name/
+        index_name="image_index",  
         store_collection_with_index=False,
         overwrite=True
     )
@@ -65,7 +63,7 @@ def process_pdf_and_query(pdf_file, user_query):
         }
     ]
 
-    # Generate text with the Qwen model
+
     text = processor.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
     image_inputs, video_inputs = process_vision_info(messages)
     inputs = processor(
@@ -77,7 +75,6 @@ def process_pdf_and_query(pdf_file, user_query):
     )
     inputs = inputs.to("cuda")
     
-    # Generate the output response
     generated_ids = model.generate(**inputs, max_new_tokens=50)
     generated_ids_trimmed = [
         out_ids[len(in_ids):] for in_ids, out_ids in zip(inputs.input_ids, generated_ids)
@@ -88,18 +85,67 @@ def process_pdf_and_query(pdf_file, user_query):
 
     return output_text[0], num_images
 
-# Define the Gradio Interface
-pdf_input = gr.File(label="Upload PDF")  # Single PDF file input
-query_input = gr.Textbox(label="Enter your query", placeholder="Ask a question about the PDF")  # User query input
-output_text = gr.Textbox(label="Model Answer")  # Output for the model's answer
-output_images = gr.Textbox(label="Number of Images in PDF")  # Output for number of images
 
-# Launch the Gradio app
+css = """
+    .duplicate-button {
+        background-color: #6272a4;
+        color: white;
+        font-weight: bold;
+        border-radius: 5px;
+        margin-top: 20px;
+        padding: 10px;
+        text-align: center;
+    }
+    .gradio-container {
+        background-color: #282a36;
+        color: #f8f8f2;
+        font-family: 'Courier New', Courier, monospace;
+        padding: 20px;
+        border-radius: 10px;
+        box-shadow: 0 4px 6px rgba(0, 0, 0, 0.1);
+    }
+"""
+
+explanation = """
+### Multimodal RAG with Image Query
+This demo showcases the **Multimodal RAG (Retriever-Augmented Generation)** model. The RAG system integrates retrieval and generation, allowing it to retrieve relevant information from a multimodal database (like PDFs with text and images) and then generate detailed responses. 
+
+We use **ColPali**, a state-of-the-art multimodal retriever, combined with the **Byaldi** library from **answer.ai**, which simplifies using ColPali. The language model used for generating answers is **Qwen/Qwen2-VL-2B-Instruct**, a powerful vision-language model capable of understanding both text and images.
+"""
+
+footer = """
+<div style="text-align: center; margin-top: 20px;">
+    <a href="https://www.linkedin.com/in/pejman-ebrahimi-4a60151a7/" target="_blank">LinkedIn</a> |
+    <a href="https://github.com/arad1367" target="_blank">GitHub</a> |
+    <a href="https://arad1367.pythonanywhere.com/" target="_blank">Live demo of my PhD defense</a> |
+    <a href="https://huggingface.co/Qwen/Qwen2-VL-2B-Instruct" target="_blank">Qwen/Qwen2-VL-2B-Instruct</a> |
+    <a href="https://github.com/AnswerDotAI/byaldi" target="_blank">Byaldi</a> |
+    <a href="https://github.com/illuin-tech/colpali" target="_blank">ColPali</a>
+    <br>
+    Made with 💖 by Pejman Ebrahimi
+</div>
+"""
+
+pdf_input = gr.File(label="Upload PDF")  
+query_input = gr.Textbox(label="Enter your query", placeholder="Ask a question about the PDF")  
+output_text = gr.Textbox(label="Model Answer")  
+output_images = gr.Textbox(label="Number of Images in PDF") 
+duplicate_button = gr.DuplicateButton(value="Duplicate Space for private use", elem_classes="duplicate-button")
+
+# Launch the Gradio app 
 demo = gr.Interface(
     fn=process_pdf_and_query, 
-    inputs=[pdf_input, query_input],  # List of inputs
-    outputs=[output_text, output_images],  # List of outputs
-    title="Multimodal RAG with Image Query - By Pejman Ebrahimi"
+    inputs=[pdf_input, query_input],  
+    outputs=[output_text, output_images],  
+    title="Multimodal RAG with Image Query - By Pejman Ebrahimi - Please like the space if it is useful",
+    theme='freddyaboulton/dracula_revamped',
+    css=css,
+    description=explanation,
+    allow_flagging="auto"
 )
 
-demo.launch(debug=True)  # Start the interface
+with demo:
+    gr.HTML(footer)
+    duplicate_button
+
+demo.launch(debug=True)