Update app.py

app.py

@@ -60,7 +60,7 @@ accelerator = Accelerator()
 model = AutoModelForCausalLM.from_pretrained(model_id, token= token, 
                                                  # torch_dtype= torch.uint8, 
                                              torch_dtype=torch.float16,
-                                              load_in_8bit=True,
+                                              load_in_4bit=True,
                                              # #  # torch_dtype=torch.fl,
                                              attn_implementation="flash_attention_2",
                                              low_cpu_mem_usage=True,
@@ -96,6 +96,8 @@ def respond(
     temperature,
     top_p,
 ):
+    # model.to(accelerator.device)
+
     messages = []
     json_obj = str_to_json(message)
     print(json_obj)