Update app.py

app.py

@@ -21,7 +21,7 @@ embedding_model = HuggingFaceEmbeddings(model_name='sentence-transformers/all-Mi
 vectordb = FAISS.load_local("faiss_index", embedding_model, allow_dangerous_deserialization=True)
 
 # --- Step 3: Load the LLM ---
-model_id = "tiiuae/falcon-7b-instruct"
+model_id = "tiiuae/falcon-3b-instruct"
 tokenizer = AutoTokenizer.from_pretrained(model_id)
 
 # ✅ Use device_map + float16 to save memory