Update app.py

app.py

@@ -11,65 +11,30 @@ DEFAULT_API_URL = "https://agents-course-unit4-scoring.hf.space"
 # --- Basic Agent Definition ---
 # ----- THIS IS WERE YOU CAN BUILD WHAT YOU WANT ------
 # ---------- MODIFICATIONS BEGIN ----------
-from transformers import AutoTokenizer, AutoModelForCausalLM
-import torch
-import os
-import re
+from transformers import pipeline
 
 class BasicAgent:
     def __init__(self):
-        print("Hybrid Agent with Mistral Model Initialized")
-
-        model_id = "mistralai/Mistral-7B-Instruct-v0.1"
-
-        self.tokenizer = AutoTokenizer.from_pretrained(model_id, token=os.getenv("HF_NEW_API_TOKEN"))
-        self.model = AutoModelForCausalLM.from_pretrained(model_id, token=os.getenv("HF_NEW_API_TOKEN"))
-        self.model.to("cpu")
-        self.model.eval()
-
-    def classify(self, question: str) -> str:
-        q = question.lower()
-        if any(x in q for x in ["youtube", ".mp3", "image", "video", "attached", ".wav"]):
-            return "media"
-        if any(x in q for x in ["|", "*", "subset", "commutative", "table", "="]):
-            return "logic"
-        return "mistral"
-
-    def handle_media(self, q: str) -> str:
-        return "I'm unable to process audio, video, or file-based questions."
-
-    def handle_logic(self, q: str) -> str:
-        q = q.lower()
-        if "not commutative" in q and "subset" in q:
-            return "a,b,c"
-        return "I couldn't solve this logic-based question."
-
-    def handle_mistral(self, question: str) -> str:
-        prompt = f"<s>[INST] {question.strip()} [/INST]"
-        inputs = self.tokenizer(prompt, return_tensors="pt").to("cpu")
-
-        with torch.no_grad():
-            outputs = self.model.generate(
-                **inputs,
-                max_new_tokens=256,
-                do_sample=True,
-                temperature=0.7,
-                top_p=0.95
-            )
-        response = self.tokenizer.decode(outputs[0], skip_special_tokens=True)
-        return response.split("[/INST]")[-1].strip()
+        print("FLAN-T5-BASE Local Agent initialized.")
+
+        self.pipeline = pipeline(
+            "text2text-generation",
+            model="google/flan-t5-base",
+            tokenizer="google/flan-t5-base",
+            device=-1  # use CPU
+        )
 
     def __call__(self, question: str) -> str:
-        qtype = self.classify(question)
-        print(f"Classified as: {qtype}")
-
-        if qtype == "media":
-            return self.handle_media(question)
-        elif qtype == "logic":
-            return self.handle_logic(question)
-        else:
-            return self.handle_mistral(question)
-# ---------- MODIFICATIONS END ----------
+        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()
+        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):
     """