Update app.py

app.py

@@ -65,7 +65,7 @@ model = AutoModelForCausalLM.from_pretrained(model_id, token= token,
                                               # load_in_8bit=True,
                                              # #  # torch_dtype=torch.fl,
                                              attn_implementation="flash_attention_2",
-                                             # low_cpu_mem_usage=True,
+                                             low_cpu_mem_usage=True,
                                              # device_map='cuda',
                                              # device_map=accelerator.device_map,