Update app.py

app.py

@@ -126,8 +126,6 @@ def load_id_answer_mapping():
     return json.loads(id_answer_mapping)
 
 
-
-
 def evaluate_uploaded_json(user_file, model_name):
     print(f"Model Name: {model_name}")
     print(f"Uploaded File: {user_file}")
@@ -155,15 +153,16 @@ def evaluate_uploaded_json(user_file, model_name):
             class_correct[question_class] += 1
             correct += 1
 
-
     subclass_data = []
-    for cls in CLASS_LIST[:-5]: 
+    subclass_result = {}  
+    for cls in CLASS_LIST[:-5]:  
         acc = class_correct[cls] / class_total[cls] if class_total[cls] > 0 else 0
         subclass_data.append({
             "Subclass": cls,
             "Accuracy": f"{acc:.2%}",
             "Correct/Total": f"{class_correct[cls]}/{class_total[cls]}"
         })
+        subclass_result[cls] = acc  
 
 
     category_data = []
@@ -176,17 +175,20 @@ def evaluate_uploaded_json(user_file, model_name):
             "Accuracy": f"{acc:.2%}",
             "Correct/Total": f"{cat_correct}/{cat_total}"
         })
+        subclass_result[category] = acc 
 
     overall_accuracy = f"{correct / total:.2%} ({correct}/{total} correct)"
 
     subclass_df = pd.DataFrame(subclass_data)
     category_df = pd.DataFrame(category_data)
 
-    save_class_accuracy_to_hf_dataset(model_name, {**{d['Subclass']: d['Accuracy'] for d in subclass_data},
-                                                   **{d['Category']: d['Accuracy'] for d in category_data}})
+
+    save_class_accuracy_to_hf_dataset(model_name, subclass_result)
 
     return overall_accuracy, category_df, subclass_df
 
+
+
 def save_class_accuracy_to_hf_dataset(model_name, class_accuracy):
 
     new_data = {"Model Name": model_name}
@@ -224,19 +226,22 @@ with demo:
         #     gr.Markdown(LLM_BENCHMARKS_TEXT, elem_classes="markdown-text")
 
         with gr.TabItem("🚀 Submit here! ", elem_id="llm-benchmark-tab-table", id=3):
-            gr.Interface(
+            with gr.Row():
+                json_file = gr.File(label="Upload JSON File")
+                model_name = gr.Textbox(label="Model Name", placeholder="Enter your model name here")
+        
+            with gr.Row():
+                overall_acc = gr.Textbox(label="Overall Accuracy")
+            
+            with gr.Row():
+                category_df = gr.Dataframe(label="Category Accuracy")
+                subclass_df = gr.Dataframe(label="Subclass Accuracy")
+        
+            json_eval_button = gr.Button("Evaluate")
+            json_eval_button.click(
                 fn=evaluate_uploaded_json,
-                inputs=[
-                    gr.File(label="Upload JSON File"),
-                    gr.Textbox(label="Model Name", placeholder="Enter your model name here")
-                ],
-                outputs=[
-                    gr.Textbox(label="Overall Accuracy"),
-                    gr.Dataframe(label="Category Accuracy"),
-                    gr.Dataframe(label="Subclass Accuracy")
-                ],
-                title="JSON Answer Evaluator",
-                description="JSON Answer Evaluator"
+                inputs=[json_file, model_name],
+                outputs=[overall_acc, category_df, subclass_df]
             )
 
     with gr.Row():