Update app.py

app.py

@@ -8,10 +8,13 @@ from langchain.docstore.document import Document
 from langchain.text_splitter import RecursiveCharacterTextSplitter
 from langchain_community.retrievers import BM25Retriever
 
-from smolagents import Tool, CodeAgent, InferenceClientModel
+from smolagents import Tool, CodeAgent
+
+from huggingface_hub.inference_api import InferenceApi  # <--- NEW import
+
 
 hf_token = os.getenv("HUGGINGFACE_API_KEY")
-print("Token from env var:", os.getenv("HUGGINGFACE_API_KEY"))
+print("Token from env var:", hf_token)
 os.environ["HUGGINGFACE_API_KEY"] = hf_token
 
 if hf_token:
@@ -19,10 +22,11 @@ if hf_token:
     print("Set HUGGINGFACE_API_KEY in env.")
 else:
     print("No HUGGINGFACE_API_KEY found in env.")
-    
+
 print("HUGGINGFACE_API_KEY in env:", "HUGGINGFACE_API_KEY" in os.environ)
 print("HUGGINGFACE_API_KEY value (masked):", os.environ.get("HUGGINGFACE_API_KEY", "")[:5] + "...")
-    
+
+
 def run_and_submit_all(profile: gr.OAuthProfile | None):
     """
     Fetches all questions, runs the agent on them, submits all answers,
@@ -94,15 +98,27 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
 
         retriever_tool = RetrieverTool(docs_processed)
 
-        # Instantiate the smolagents CodeAgent
+        # Initialize HF Inference API client with your token
+        inference = InferenceApi(repo_id="Qwen/Qwen2.5-VL-7B-Instruct", token=hf_token)
+
+        # Wrap inference client in a callable class for smolagents compatibility
+        class HuggingFaceInferenceWrapper:
+            def __init__(self, client):
+                self.client = client
+
+            def __call__(self, prompt: str):
+                response = self.client(inputs=prompt)
+                if isinstance(response, dict):
+                    # Usually HF text generation returns {'generated_text': "..."}
+                    return response.get("generated_text") or str(response)
+                return str(response)
+
         agent = CodeAgent(
             tools=[retriever_tool],
-            model=InferenceClientModel(model_id="Qwen/Qwen2.5-VL-7B-Instruct"),
+            model=HuggingFaceInferenceWrapper(inference),
             max_steps=4,
             verbosity_level=2,
             stream_outputs=True,
-            
-            
         )
 
     except Exception as e:
@@ -245,9 +261,6 @@ if __name__ == "__main__":
         print(f"   Repo URL: https://huggingface.co/spaces/{space_id_startup}")
         print(f"   Repo Tree URL: https://huggingface.co/spaces/{space_id_startup}/tree/main")
     else:
-        print("ℹ️  SPACE_ID environment variable not found (running locally?). Repo URL cannot be determined.")
-
-    print("-" * (60 + len(" App Starting ")) + "\n")
+        print("ℹ️  SPACE_ID environment variable not found (running locally?).")
 
-    print("Launching Gradio Interface for Basic Agent Evaluation...")
-    demo.launch(debug=True, share=False)
+    demo.launch(debug=True)