Update app.py

app.py

@@ -55,7 +55,7 @@ tokenizer = AutoTokenizer.from_pretrained(
     , token= token,)
 
 
-accelerator = Accelerator()
+# accelerator = Accelerator()
 
 model = AutoModelForCausalLM.from_pretrained(model_id, token= token, 
                                                  # torch_dtype= torch.uint8, 
@@ -71,7 +71,7 @@ model = AutoModelForCausalLM.from_pretrained(model_id, token= token,
 
 
 # 
-model = accelerator.prepare(model)
+# model = accelerator.prepare(model)
 
 
 # device_map = infer_auto_device_map(model, max_memory={0: "79GB", "cpu":"65GB" })
@@ -102,7 +102,7 @@ def respond(
     
     messages= json_obj
 
-    input_ids = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt").to(accelerator.device)
+    input_ids = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt").to('cuda') #.to(accelerator.device)
     input_ids2 = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True, return_tensors="pt") #.to('cuda')
     print(f"Converted input_ids dtype: {input_ids.dtype}")
     input_str= str(input_ids2)