Update app.py

app.py

@@ -30,8 +30,8 @@ class BasicAgent:
     def __init__(self):
         print("BasicAgent initialized.")
         sambanova_api_key = os.getenv("SAMBANOVA_API_KEY")
-        model = OpenAIServerModel(
-            model_id="Meta-Llama-3.1-8B-Instruct",
+        self.model = OpenAIServerModel(
+            model_id="Meta-Llama-3.3-70B-Instruct",
             api_base="https://api.sambanova.ai/v1",
             api_key=sambanova_api_key,
         )
@@ -41,21 +41,31 @@ class BasicAgent:
             PythonInterpreterTool(),
             get_current_date_time,
         ]
-        self.agent = ToolCallingAgent(
-            tools=tools,
-            model=model,
-            max_steps=8,
-        )
+        self.tools = tools
 
     def __call__(self, question: str) -> str:
         print(f"Agent received question (first 50 chars): {question[:50]}")
-        try:
-            answer = self.agent.run(question)
-            print(f"Agent answer: {str(answer)[:100]}")
-            return str(answer)
-        except Exception as e:
-            print(f"Agent error: {e}")
-            return f"Error: {e}"
+        # Retry up to 3 times with backoff on rate limit errors
+        for attempt in range(3):
+            try:
+                agent = ToolCallingAgent(
+                    tools=self.tools,
+                    model=self.model,
+                    max_steps=8,
+                )
+                answer = agent.run(question)
+                print(f"Agent answer: {str(answer)[:100]}")
+                return str(answer)
+            except Exception as e:
+                err = str(e)
+                if "429" in err or "rate_limit" in err.lower() or "rate limit" in err.lower():
+                    wait = 60 * (attempt + 1)
+                    print(f"Rate limit hit, waiting {wait}s before retry {attempt+1}/3...")
+                    time.sleep(wait)
+                else:
+                    print(f"Agent error: {e}")
+                    return f"Error: {e}"
+        return "Error: Rate limit exceeded after retries"
 
 
 # --- run_and_submit_all function (Keep as is) ---
@@ -111,8 +121,8 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
         except Exception as e:
             print(f"Error running agent on task {task_id}: {e}")
             results_log.append({"Task ID": task_id, "Question": question_text, "Submitted Answer": f"AGENT ERROR: {e}"})
-        # Small pause between questions to avoid rate limits
-        time.sleep(2)
+        # Pause between questions to avoid per-minute rate limits
+        time.sleep(3)
 
     if not answers_payload:
         print("Agent did not produce any answers to submit.")