Update src/streamlit_app.py

src/streamlit_app.py

@@ -23,6 +23,7 @@ with st.spinner(text="Loading model, please wait...", show_time=True):
     model, tokenizer = get_model()
 
 
+@torch.inference_mode()
 def translate(src_lang: str, tgt_lang: str, query: str) -> str:
     text = f"Translate {src_lang} to {tgt_lang}.\nQuery: {query}\nAnswer:"
     tokens = tokenizer(text, return_tensors="pt").to(DEVICE)