Update app.py

app.py

@@ -1,13 +1,8 @@
 import os
 import gradio as gr
 import pandas as pd
-import requests
 
-# --- Constants ---
-DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
-HF_MODEL = "bigscience/bloom-560m"
-HF_API_URL = f"https://api-inference.huggingface.co/models/{HF_MODEL}"
-HF_API_TOKEN = os.getenv("HF_API_TOKEN")
+from transformers import pipeline
 
 # --- Dummy Agent ---
 class DummyAgent:
@@ -17,30 +12,23 @@ class DummyAgent:
     def __call__(self, question: str) -> str:
         return "Error: Could not generate answer."
 
-# --- Hugging Face Inference API Agent ---
-class HFInferenceAgent:
+# --- Local LLM Agent using distilgpt2 ---
+class LocalLLMAgent:
     def __init__(self):
-        if not HF_API_TOKEN:
-            raise ValueError("HF_API_TOKEN environment variable is not set.")
-        self.headers = {"Authorization": f"Bearer {HF_API_TOKEN}"}
+        print("Loading local model distilgpt2...")
+        self.generator = pipeline("text-generation", model="distilgpt2")
 
     def __call__(self, question: str) -> str:
-        payload = {
-            "inputs": question,
-            "parameters": {"max_new_tokens": 100, "return_full_text": False},
-            "options": {"use_cache": False}
-        }
         try:
-            response = requests.post(HF_API_URL, headers=self.headers, json=payload, timeout=30)
-            response.raise_for_status()
-            result = response.json()
-            return result[0]['generated_text'].strip()
+            outputs = self.generator(question, max_length=50, do_sample=True, top_p=0.9, temperature=0.8)
+            return outputs[0]['generated_text'].strip()
         except Exception as e:
-            print(f"[HF Inference API Error] {e}")
+            print(f"[LocalLLMAgent Error] {e}")
             return "Error: Could not generate answer."
 
 # --- Main Run & Submit Function ---
 def run_and_submit_all(profile: gr.OAuthProfile | None):
+    DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
     space_id = os.getenv("SPACE_ID", "XeroDN/final_assignment")
 
     if profile:
@@ -52,15 +40,16 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
     questions_url = f"{DEFAULT_API_URL}/questions"
     submit_url = f"{DEFAULT_API_URL}/submit"
 
-    # Initialize agent
+    # Initialize agent - use local LLM agent (no API calls)
     try:
-        agent = HFInferenceAgent()
+        agent = LocalLLMAgent()
     except Exception as e:
         print("Falling back to DummyAgent:", e)
         agent = DummyAgent()
 
     agent_code = f"https://huggingface.co/spaces/{space_id}/tree/main"
 
+    import requests
     try:
         response = requests.get(questions_url, timeout=15)
         response.raise_for_status()
@@ -113,7 +102,7 @@ def run_and_submit_all(profile: gr.OAuthProfile | None):
 
 # --- Gradio UI ---
 with gr.Blocks() as demo:
-    gr.Markdown("## 🤖 GAIA Benchmark Agent Submission (Hugging Face Inference API)")
+    gr.Markdown("## 🤖 GAIA Benchmark Agent Submission (Local LLM, no API key)")
     gr.Markdown("Click below to log in and run your agent on the GAIA benchmark.")
 
     gr.LoginButton()