Update app.py

app.py

@@ -81,9 +81,9 @@ model = AutoModelForCausalLM.from_pretrained(
     torch_dtype="auto",
     attn_implementation="flash_attention_2",
     quantization_config=quantization_config,
-    # low_cpu_mem_usage=True,
-    # device_map="auto"     
-).to(device)
+    low_cpu_mem_usage=True,
+    device_map="auto"     
+)
 model.eval()
 
 basic_model_config = './xcodec_mini_infer/final_ckpt/config.yaml'