BentoML

I have added device in the pipeline example
```device = torch.device("cuda:0" if torch.cuda.is_available() else "cpu")
pipeline(task="text2text-generation", model=model, tokenizer=tokenizer,device=device)```
Further i call runner for inference
```runner = bentoml.transformers.get("model:latest").to_runner()
service = bentoml.Service("model", runners=[runner])
@service.api():
runer.run("its an example")```

that is not a valid service definition no?

Yes, this is not a valid, i changed due to privacy. But it has essential elements