Update app.py

app.py

@@ -10,31 +10,35 @@ DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 
 # --- Basic Agent Definition ---
 # ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
-from transformers import pipeline
+from transformers import pipeline, AutoTokenizer, AutoModelForCausalLM
 
 class BasicAgent:
     def __init__(self):
-        print("FLAN-T5-SMALL Local Agent initialized.")
+        print("Mistral Local Agent initialized.")
 
+        model_id = "mistralai/Mistral-7B-Instruct-v0.1"
+
+        self.tokenizer = AutoTokenizer.from_pretrained(model_id)
+        self.model = AutoModelForCausalLM.from_pretrained(model_id)
         self.pipeline = pipeline(
-            "text2text-generation",
-            model="google/flan-t5-small",
-            tokenizer="google/flan-t5-small",
-            device=-1 
+            "text-generation",
+            model=self.model,
+            tokenizer=self.tokenizer,
+            device=-1  # CPU
         )
 
     def __call__(self, question: str) -> str:
         print(f"Agent received question (first 50 chars): {question[:50]}...")
 
         try:
-            prompt = f"Answer the following question:\n{question.strip()}"
-            result = self.pipeline(prompt, max_new_tokens=128, temperature=0.5)
-            answer = result[0]["generated_text"]
-            return answer.strip()
+            prompt = f"<s>[INST] {question.strip()} [/INST]"
+            result = self.pipeline(prompt, max_new_tokens=256, temperature=0.7)
+            return result[0]["generated_text"].split("[/INST]")[-1].strip()
         except Exception as e:
             print(f"❌ Error during model inference: {e}")
             return f"❌ Model Error: {str(e)}"
 
+
 def run_and_submit_all( profile: gr.OAuthProfile | None):
     """
     Fetches all questions, runs the BasicAgent on them, submits all answers,