Update app.py

app.py

@@ -29,7 +29,7 @@ else:
     faiss.write_index(index, "faiss.index")
 
 # --- Inference API client ---
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta", token=os.getenv("HF_TOKEN"))
+client = InferenceClient("mistralai/Mistral-7B-Instruct-v0.3", token=os.getenv("HF_TOKEN"))
 
 def respond(message, history):
     query_embedding = embedding_model.encode([message], convert_to_numpy=True)