Update app.py

app.py

@@ -230,8 +230,8 @@ def load_model():
         model = AutoModelForSeq2SeqLM.from_pretrained(
             MODEL_NAME,
             torch_dtype=torch.float32,
-            low_cpu_mem_usage=True,
         )
+        model = model.to("cpu")
         model.eval()
         logging.info(f"Loaded {MODEL_NAME}")
         return tokenizer, model