Update app.py

app.py

@@ -106,33 +106,36 @@ def load_responses():
     with open("result/GSM8K/gsm_result_cot.json", "r") as gsm_cot_file:
         gsm_cot_responses = json.load(gsm_cot_file)
 
-    # with open("result/MMLU/mmlu_result.json", "r") as mmlu_file:
-    #     mmlu_responses = json.load(mmlu_file)
+    with open("result/MMLU/mmlu_result.json", "r") as mmlu_file:
+        mmlu_responses = json.load(mmlu_file)
 
-    # with open("result/MMLU/mmlu_result_cot.json", "r") as mmlu_cot_file:
-    #     mmlu_cot_responses = json.load(mmlu_cot_file)
+    with open("result/MMLU/mmlu_result_cot.json", "r") as mmlu_cot_file:
+        mmlu_cot_responses = json.load(mmlu_cot_file)
     
-    return math_responses, math_cot_responses, gsm_responses, gsm_cot_responses#, mmlu_responses
+    return math_responses, math_cot_responses, gsm_responses, gsm_cot_responses, mmlu_responses, mmlu_cot_responses
 
 def load_questions(math, gsm):
     math_questions = []
     gsm_questions = []
-    # mmlu_questions = []
+    mmlu_questions = []
     for i in range(100):
         math_questions.append(f"{i+1}: " + math[i]["question"][:128] + "...")
         gsm_questions.append(f"{i+1}: " + gsm[i]["question"][:128] + "...")
-        # mmlu_questions.append(f"{i+1}: " + mmlu[i]["question"][:128] + "...")
+        mmlu_questions.append(f"{i+1}: " + mmlu[i]["question"][:128] + "...")
 
-    return math_questions, gsm_questions#, mmlu_questions
+    return math_questions, gsm_questions, mmlu_questions
 
-math_result, math_cot_result, gsm_result, gsm_cot_result = load_responses()
+math_result, math_cot_result, gsm_result, gsm_cot_result, mmlu_result, mmlu_cot_result = load_responses()
 
-math_questions, gsm_questions = load_questions(math_result, gsm_result)
+math_questions, gsm_questions, mmlu_questions = load_questions(math_result, gsm_result, mmlu_result)
 
 math_question_selector_map = build_question_selector_map(math_result)
 math_cot_question_selector_map = build_question_selector_map(math_cot_result)
 gsm_question_selector_map = build_question_selector_map(gsm_result)
 gsm_cot_question_selector_map = build_question_selector_map(gsm_cot_result)
+mmlu_question_selector_map = build_question_selector_map(mmlu_result)
+mmlu_cot_question_selector_map = build_question_selector_map(mmlu_cot_result)
+
 
 TITLE = """<h1 align="center">LLM Agora 🗣️🏦</h1>"""
 
@@ -317,7 +320,7 @@ with gr.Blocks() as demo:
 
         with gr.Tab("MMLU"):
             mmlu_cot = gr.Checkbox(label="CoT", info="If you want to see CoT result, please check the box.")
-            # mmlu_question_list = gr.Dropdown(mmlu_questions, value=callable(mmlu_questions), label="MMLU Question", every=0.1)
+            mmlu_question_list = gr.Dropdown(mmlu_questions, value=callable(mmlu_questions), label="MMLU Question", every=0.1)
             
             with gr.Column():
                 with gr.Row(elem_id="model1_response"):
@@ -336,18 +339,18 @@ with gr.Blocks() as demo:
                     mmlu_model3_output3 = gr.Textbox(label="Orca🐬's 3️⃣rd response")
 
             gr.HTML("""<h1 align="center"> The result of MMLU </h1>""")
-            # gr.Image(value="result/MMLU/MMLU_performance.png")
-
-        # mmlu_cot.select(
-        #     mmlu_display_question_answer,
-        #     [mmlu_question_list, mmlu_cot],
-        #     [mmlu_model1_output1, mmlu_model2_output1, mmlu_model3_output1, mmlu_summarization_text1, mmlu_model1_output2, mmlu_model2_output2, mmlu_model3_output2, mmlu_summarization_text2, mmlu_model1_output3, mmlu_model2_output3, mmlu_model3_output3]
-        # )
-        # mmlu_question_list.change(
-        #     mmlu_display_question_answer,
-        #     [mmlu_question_list, mmlu_cot],
-        #     [mmlu_model1_output1, mmlu_model2_output1, mmlu_model3_output1, mmlu_summarization_text1, mmlu_model1_output2, mmlu_model2_output2, mmlu_model3_output2, mmlu_summarization_text2, mmlu_model1_output3, mmlu_model2_output3, mmlu_model3_output3]
-        # )
+            gr.HTML("""<p align="center"><img src="https://github.com/composable-models/llm_multiagent_debate/assets/80087878/963571aa-228b-4d73-9082-5f528552383e"""")
+
+        mmlu_cot.select(
+            mmlu_display_question_answer,
+            [mmlu_question_list, mmlu_cot],
+            [mmlu_model1_output1, mmlu_model2_output1, mmlu_model3_output1, mmlu_summarization_text1, mmlu_model1_output2, mmlu_model2_output2, mmlu_model3_output2, mmlu_summarization_text2, mmlu_model1_output3, mmlu_model2_output3, mmlu_model3_output3]
+        )
+        mmlu_question_list.change(
+            mmlu_display_question_answer,
+            [mmlu_question_list, mmlu_cot],
+            [mmlu_model1_output1, mmlu_model2_output1, mmlu_model3_output1, mmlu_summarization_text1, mmlu_model1_output2, mmlu_model2_output2, mmlu_model3_output2, mmlu_summarization_text2, mmlu_model1_output3, mmlu_model2_output3, mmlu_model3_output3]
+        )
 
         with gr.Accordion("※ Specific information about LLM Agora", open=False):
             gr.Markdown(SPECIFIC_INFORMATION)