Revertir la inicialización del modelo a HfApiModel y eliminar la importación de InferenceClient

app.py

@@ -1,5 +1,4 @@
 from smolagents import CodeAgent, HfApiModel, load_tool, tool
-from huggingface_hub import InferenceClient
 import datetime
 import requests
 import pytz
@@ -80,16 +79,12 @@ web_searcher = DuckDuckGoSearchTool(max_results=5)
 # If the agent does not answer, the model is overloaded, please use another model or the following Hugging Face Endpoint that also contains qwen2.5 coder:
 # model_id='https://pflgm2locj2t89co.us-east-1.aws.endpoints.huggingface.cloud' 
 
-# model = HfApiModel(
-#     max_tokens=2096,
-#     temperature=0.5,
-#     model_id='Qwen/Qwen2.5-Coder-32B-Instruct',
-#     custom_role_conversions=None,
-#     token=HF_TOKEN
-# )
-model = InferenceClient(
-    model="Qwen/Qwen2.5-Coder-32B-Instruct",
-    api_key=HF_TOKEN
+model = HfApiModel(
+    max_tokens=2096,
+    temperature=0.5,
+    model_id='Qwen/Qwen2.5-Coder-32B-Instruct',
+    custom_role_conversions=None,
+    token=HF_TOKEN
 )