Update app.py

app.py

@@ -31,7 +31,7 @@ def init():
     model = AutoModelForCausalLM.from_pretrained(
             model_id, token=huggingface_token ,torch_dtype=dtype,device_map=device
         )
-    text_generator = pipeline("text-generation", model=model, tokenizer=tokenizer,torch_dtype=dtype,device_map=device) #pipeline has not to(device)
+    text_generator = pipeline("text-generation", model=model, tokenizer=tokenizer,torch_dtype=dtype,device_map=device,device=0 ) #pipeline has not to(device)
     
     if next(model.parameters()).is_cuda:
         print("The model is on a GPU")