Update Aug 22 12:58

app.py

@@ -50,7 +50,7 @@ def s(input_string):
 
 def optimize(objective_function, iteration_input, progress=gr.Progress()):
 
-    print(objective_function)
+    # print(objective_function)
 
     # Variable setup
     Current_BEST = torch.tensor( -1e10 )   # Some arbitrary very small number
@@ -153,8 +153,8 @@ def optimize(objective_function, iteration_input, progress=gr.Progress()):
         
         # (ii) Convergence tracking (assuming the best Y is to be maximized)
         # if Current_BEST != -1e10:
-        print(Current_BEST)
-        print(convergence)
+        # print(Current_BEST)
+        # print(convergence)
         convergence.append(Current_BEST.abs())
         time_conv.append(time.time() - START_TIME)
 
@@ -258,7 +258,7 @@ def create_convergence_plot(objective_function, iteration_input, time_conv, conv
 
     
     ax.set_xlabel('Time (seconds)')
-    ax.set_ylabel('Objective Value')
+    ax.set_ylabel('Objective Value (Minimization)')
     ax.set_title('Convergence Plot for {t} iterations'.format(t=iteration_input))
     # ax.legend()
 
@@ -284,7 +284,7 @@ def create_convergence_plot(objective_function, iteration_input, time_conv, conv
         ax.axhline(y=2650, color='red', linestyle='--', label='Optimal Value')
         
     elif objective_function=="WeldedBeam.png":
-        ax.axhline(y=6, color='red', linestyle='--', label='Optimal Value')  
+        ax.axhline(y=3.3, color='red', linestyle='--', label='Optimal Value')  
 
     elif objective_function=="Car.png":
         ax.axhline(y=25, color='red', linestyle='--', label='Optimal Value')  
@@ -367,7 +367,7 @@ def submit_action(objective_function_choices, iteration_input):
 def clear_output():
     # print(gallery.selected_index)
     
-    return gr.update(value=[], selected=None),  None, 15, gr.Markdown(""), 'Test_formulation_default.png'
+    return gr.update(value=[], selected=None),  None, 15, gr.Markdown(""), 'Formulation_default.png'
 
 def reset_gallery():
     return gr.update(value=image_paths)
@@ -378,14 +378,32 @@ with gr.Blocks() as demo:
     gr.HTML(
         """
         <div style="text-align: center;">
-            <h1>Pre-trained Transformer for Constrained Bayesian Optimization</h1>
-            <h4>Paper: <a href="https://arxiv.org/abs/2404.04495">
+            <p style="text-align: center; font-size:30px;"><b>
+            Constrained Bayesian Optimization with Pre-trained Transformers
+            </b></p>
+            
+            <p style="text-align: center; font-size:18px;"><b>
+            Paper: <a href="https://arxiv.org/abs/2404.04495">
             Fast and Accurate Bayesian Optimization with Pre-trained Transformers for Constrained Engineering Problems</a> 
-            </h4>
+            </b></p>
+
+            <p style="text-align: left;font-size:18px;">
+            Explore our interactive demo that uses PFN (Prior-Data Fitted Networks) for solving constrained Bayesian optimization problems! 
+            </p>
+            
+            <p style="text-align: left;font-size:24px;"><b>
+	    Get Started:
+	    </b> </p>
+
+
+            <p style="text-align: left;font-size:18px;">
+             <ol style="text-align: left;font-size:18px;text-indent: 30px;">
+              <li> <b>Select a Problem:</b> Click on an image from the problem gallery to choose your objective function. </li>
+              <li> <b>Set Iterations:</b> Adjust the slider to set the number of iterations for the optimization process. </li>
+              <li> <b>Run Optimization:</b> Click "Submit" to start the optimization. Use "Clear" if you need to reselect your parameters. </li> 
+            </ol> 
+            </p>
 
-            <p style="text-align: left;">This is a demo for Bayesian Optimization using PFN (Prior-Data Fitted Networks). 
-            Select your objective function by clicking on one of the check boxes below, then enter the iteration number to run the optimization process. 
-            The results will be visualized in the radar chart and convergence plot.</p>
             
             
             
@@ -393,6 +411,26 @@ with gr.Blocks() as demo:
         </div>
         """
     )
+    
+    gr.HTML(
+    
+            	"""
+            	    <p style="text-align: left;font-size:24px;"><b>
+		    Result Display:
+		    </b> </p>
+		    
+		    <p style="text-align: left;font-size:18px;">
+		    <ol style="text-align: left;font-size:18px;text-indent: 30px;">
+		    <li> <b>Panel Display:</b> Shows the problem formulation and the optimization results. </li>
+		    <li> <b>Convergence Plot:</b> Visualizes the best observed objective against the algorithm's runtime over the chosen iterations. </li>
+		    	<ul>
+		    	<li> <b>PFN-CBO:</b> Displays results from real-time optimization. </li>
+		    	<li> <b>GP-CBO:</b> Provides pre-computed data from our past experiments, as GP real-time runs are impractical for a demo. </li>
+		    	</ul>
+		    </ol> 
+		    </p>
+            	"""
+            )
 
     
     with gr.Row():
@@ -405,7 +443,7 @@ with gr.Blocks() as demo:
                 gr.Markdown("## Select a problem (objective): ")
                 img_key = gr.Markdown(value="", visible=False)
             
-            gallery = gr.Gallery(value=image_paths, label="Objective Functions", 
+            gallery = gr.Gallery(value=image_paths, label="Objectives", 
                                  # height = 450, 
                                  object_fit='contain',
                                  columns=3, rows=3, elem_id="gallery")
@@ -421,20 +459,54 @@ with gr.Blocks() as demo:
 
         with gr.Column():
             # gr.Markdown("# Outputs: ")
-            gr.Markdown("## Problem Formulation: ")
-            formulation = gr.Image(value='Formulation_default.png', height=150)
-            gr.Markdown("## Results: ")
-            gr.Markdown("The graph will plot the best observed data v.s. the time for the algorithm to run up until the iteration. The PFN-CBO shows the result of the realtime optimization running in the backend while the GP-CBO shows the stored data from our previous experiments since running GP-CBO will take longer time.")
+            gr.Markdown("""
+		        ## Convergence Plot: 
+		        """)
+            
             convergence_plot = gr.Plot(label="Convergence Plot")
+            
+
+            gr.Markdown("")
+            gr.Markdown("## Problem formulation: ")
+            formulation = gr.Image(value='Formulation_default.png', label="Eq")
+            
+            
 
 
 
     def handle_select(evt: gr.SelectData):
         selected_image = evt.value
         key = evt.value['image']['orig_name']
-        formulation = 'Test_formulation.png'
-        print('here')
-        print(key)
+
+        if key=="CantileverBeam.png":
+            formulation = 'Cantilever_formulation.png'
+    
+        elif key=="CompressionSpring.png":
+            formulation = 'Compressed_Formulation.png'
+    
+        elif key=="HeatExchanger.png":
+            formulation = 'Heat_Formulation.png'
+            
+        elif key=="Reinforcement.png":
+            formulation = 'Reinforce_Formulation.png'
+            
+        elif key=="PressureVessel.png":
+            formulation = 'Pressure_Formulation.png'
+            
+        elif key=="SpeedReducer.png":
+            formulation = 'Speed_Formulation.png'
+            
+        elif key=="WeldedBeam.png":
+            formulation = 'Welded_Formulation.png'  
+    
+        elif key=="Car.png":
+            formulation = 'Car_Formulation_2.png'
+
+
+        
+        # formulation = 'Test_formulation.png'
+        # print('here')
+        # print(key)
 
         return key, formulation
         
@@ -465,4 +537,4 @@ with gr.Blocks() as demo:
 
     
 
-demo.launch()
+demo.launch(share=True, server_port=7959)