Update app.py

app.py

@@ -15,6 +15,8 @@ import torch
 model_dir = "TheBloke_Wizard-Vicuna-13B-GPTQ"  # Path to downloaded model
 config = ExLlamaV2Config()
 config.model_dir = model_dir
+config.device_map = "cpu"
+config.no_flash_attn = True  # Disable flash attention
 config.prepare()
 
 # Load model