Update app.py

app.py

@@ -81,104 +81,100 @@ with gr.Blocks(fill_height=True) as demo:
     gr.Markdown("**Important note**: This model is not made for chatting, the chatty IDEFICS2 will be released in the upcoming days. **This model is very strong on various tasks, including visual question answering, document retrieval and more.**")
     gr.Markdown("Learn more about IDEFICS2 in this [blog post](https://huggingface.co/blog/idefics2).")
 
-    with gr.Row():
-      with gr.Column():
+
+    with gr.Column():
         image_input = gr.Image(label="Upload your Image", type="pil")
         query_input = gr.Textbox(label="Prompt")
         submit_btn = gr.Button("Submit")
-    
-      with gr.Column():
-
         output = gr.Textbox(label="Output")
-
-    with gr.Accordion():
-      # Hyper-parameters for generation
-      max_new_tokens = gr.Slider(
-          minimum=8,
-          maximum=1024,
-          value=512,
-          step=1,
-          interactive=True,
-          label="Maximum number of new tokens to generate",
-      )
-      repetition_penalty = gr.Slider(
-          minimum=0.01,
-          maximum=5.0,
-          value=1.2,
-          step=0.01,
-          interactive=True,
-          label="Repetition penalty",
-          info="1.0 is equivalent to no penalty",
-      )
-      temperature = gr.Slider(
-          minimum=0.0,
-          maximum=5.0,
-          value=0.4,
-          step=0.1,
-          interactive=True,
-          label="Sampling temperature",
-          info="Higher values will produce more diverse outputs.",
-      )
-      top_p = gr.Slider(
-          minimum=0.01,
-          maximum=0.99,
-          value=0.8,
-          step=0.01,
-          interactive=True,
-          label="Top P",
-          info="Higher values is equivalent to sampling more low-probability tokens.",
-      )
-      decoding_strategy = gr.Radio(
-          [
-              "Greedy",
-              "Top P Sampling",
-          ],
-          value="Greedy",
-          label="Decoding strategy",
-          interactive=True,
-          info="Higher values is equivalent to sampling more low-probability tokens.",
-      )
-      decoding_strategy.change(
-          fn=lambda selection: gr.Slider(
-              visible=(
-                  selection in ["contrastive_sampling", "beam_sampling", "Top P Sampling", "sampling_top_k"]
-              )
-          ),
-          inputs=decoding_strategy,
-          outputs=temperature,
-      )
-
-      decoding_strategy.change(
-          fn=lambda selection: gr.Slider(
-              visible=(
-                  selection in ["contrastive_sampling", "beam_sampling", "Top P Sampling", "sampling_top_k"]
-              )
-          ),
-          inputs=decoding_strategy,
-          outputs=repetition_penalty,
-      )
-      decoding_strategy.change(
-          fn=lambda selection: gr.Slider(visible=(selection in ["Top P Sampling"])),
-          inputs=decoding_strategy,
-          outputs=top_p,
-      )
-      examples=[["./example_images/docvqa_example.png", "How many items are sold?", "Greedy", 0.4, 512, 1.2, 0.8],
-                ["./example_images/s2w_example.png", "What is this UI about?", "Greedy", 0.4, 512, 1.2, 0.8],
-                ["./example_images/example_images_travel_tips.jpg", "I want to go somewhere similar to the one in the photo. Give me destinations and travel tips.",  0.4, 512, 1.2, 0.8],
-                ["./example_images/chicken_on_money.png", "Can you tell me a very short story based on this image?", 0.4, 512, 1.2, 0.8],
-                ["./example_images/baklava.png", "Where is this pastry from?", 0.4, 512, 1.2, 0.8],
-                ["./example_images/dummy_pdf.png", "How much percent is the order status?", 0.4, 512, 1.2, 0.8],
-                ["./example_images/art_critic.png", "As an art critic AI assistant, could you describe this painting in details and make a thorough critic?.", 0.4, 512, 1.2, 0.8]]
-      gr.Examples(
-                examples = examples,
-                inputs=[image_input, query_input, decoding_strategy, temperature, 
-                                                      max_new_tokens, repetition_penalty, top_p],
-                outputs=output,
-                fn=model_inference
-            )
-      submit_btn.click(model_inference, inputs = [image_input, query_input, decoding_strategy, temperature, 
-                                                  max_new_tokens, repetition_penalty, top_p],
-                       outputs=output)
+        examples=[["./example_images/docvqa_example.png", "How many items are sold?", "Greedy", 0.4, 512, 1.2, 0.8],
+                    ["./example_images/s2w_example.png", "What is this UI about?", "Greedy", 0.4, 512, 1.2, 0.8],
+                    ["./example_images/example_images_travel_tips.jpg", "I want to go somewhere similar to the one in the photo. Give me destinations and travel tips.",  0.4, 512, 1.2, 0.8],
+                    ["./example_images/chicken_on_money.png", "Can you tell me a very short story based on this image?", 0.4, 512, 1.2, 0.8],
+                    ["./example_images/baklava.png", "Where is this pastry from?", 0.4, 512, 1.2, 0.8],
+                    ["./example_images/dummy_pdf.png", "How much percent is the order status?", 0.4, 512, 1.2, 0.8],
+                    ["./example_images/art_critic.png", "As an art critic AI assistant, could you describe this painting in details and make a thorough critic?.", 0.4, 512, 1.2, 0.8]]
+        gr.Examples(
+                    examples = examples,
+                    inputs=[image_input, query_input, decoding_strategy, temperature, 
+                                                          max_new_tokens, repetition_penalty, top_p],
+                    outputs=output,
+                    fn=model_inference
+                )
+        with gr.Accordion():
+          # Hyper-parameters for generation
+          max_new_tokens = gr.Slider(
+              minimum=8,
+              maximum=1024,
+              value=512,
+              step=1,
+              interactive=True,
+              label="Maximum number of new tokens to generate",
+          )
+          repetition_penalty = gr.Slider(
+              minimum=0.01,
+              maximum=5.0,
+              value=1.2,
+              step=0.01,
+              interactive=True,
+              label="Repetition penalty",
+              info="1.0 is equivalent to no penalty",
+          )
+          temperature = gr.Slider(
+              minimum=0.0,
+              maximum=5.0,
+              value=0.4,
+              step=0.1,
+              interactive=True,
+              label="Sampling temperature",
+              info="Higher values will produce more diverse outputs.",
+          )
+          top_p = gr.Slider(
+              minimum=0.01,
+              maximum=0.99,
+              value=0.8,
+              step=0.01,
+              interactive=True,
+              label="Top P",
+              info="Higher values is equivalent to sampling more low-probability tokens.",
+          )
+          decoding_strategy = gr.Radio(
+              [
+                  "Greedy",
+                  "Top P Sampling",
+              ],
+              value="Greedy",
+              label="Decoding strategy",
+              interactive=True,
+              info="Higher values is equivalent to sampling more low-probability tokens.",
+          )
+          decoding_strategy.change(
+              fn=lambda selection: gr.Slider(
+                  visible=(
+                      selection in ["contrastive_sampling", "beam_sampling", "Top P Sampling", "sampling_top_k"]
+                  )
+              ),
+              inputs=decoding_strategy,
+              outputs=temperature,
+          )
+    
+          decoding_strategy.change(
+              fn=lambda selection: gr.Slider(
+                  visible=(
+                      selection in ["contrastive_sampling", "beam_sampling", "Top P Sampling", "sampling_top_k"]
+                  )
+              ),
+              inputs=decoding_strategy,
+              outputs=repetition_penalty,
+          )
+          decoding_strategy.change(
+              fn=lambda selection: gr.Slider(visible=(selection in ["Top P Sampling"])),
+              inputs=decoding_strategy,
+              outputs=top_p,
+          )
+          
+        submit_btn.click(model_inference, inputs = [image_input, query_input, decoding_strategy, temperature, 
+                                                      max_new_tokens, repetition_penalty, top_p], outputs=output)
       
 
 demo.launch(debug=True)