Update app.py

app.py

@@ -5,7 +5,12 @@ import inspect
 import pandas as pd
 
 
-from smolagents import CodeAgent, WebSearchTool, InferenceClientModel
+from smolagents import CodeAgent, WebSearchTool#, InferenceClientModel
+
+# Load model directly
+from transformers import AutoTokenizer, AutoModelForCausalLM
+
+
 
 # (Keep Constants as is)
 # --- Constants ---
@@ -20,7 +25,10 @@ class BasicAgent:
         print(f"Agent received question (first 50 chars): {question[:50]}...")
         fixed_answer = "This is a default answer."
 
-        model = InferenceClientModel()
+        #model = InferenceClientModel()
+        tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-Coder-32B-Instruct")
+        model = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-Coder-32B-Instruct")
+        
         agent = CodeAgent(tools=[WebSearchTool()], model=model, stream_outputs=True)
         custom_answer = agent.run("How many seconds would it take for a leopard at full speed to run through Pont des Arts?")