using transformer model

app.py

@@ -1,4 +1,5 @@
 from smolagents import CodeAgent,DuckDuckGoSearchTool, HfApiModel,load_tool,tool
+from smolagents.models import TransformersModel
 import datetime
 import requests
 import pytz
@@ -97,9 +98,15 @@ final_answer = FinalAnswerTool()
 # model_id='https://wxknx1kg971u7k1n.us-east-1.aws.endpoints.huggingface.cloud',# it is possible that this model may be overloaded
 # custom_role_conversions=None,
 # )
-model = LiteLLMModel(
-        model_id="ollama/qwen2.5-coder:32b"
-        )
+
+engine = TransformersModel(
+    model_id="meta-llama/Llama-3.2-3B-Instruct",
+    device_map="cuda",  # or "cpu" if you don't have a GPU
+    trust_remote_code=True,  # if your model requires remote code execution
+    torch_dtype='auto',
+    max_new_tokens=5000,
+    temperature=0.5
+)
 
 # # Import tool from Hub
 # image_generation_tool = load_tool("agents-course/text-to-image", trust_remote_code=True)
@@ -108,7 +115,7 @@ with open("prompts.yaml", 'r') as stream:
     prompt_templates = yaml.safe_load(stream)
     
 agent = CodeAgent(
-    model=model,
+    model=engine,
     tools=[
         final_answer, 
         get_current_time_in_timezone,