Add `n_gpu_layers` arg to langchain.llms.LlamaCpp

### Feature request

`langchain.llms.LlamaCpp` wraps around `llama_cpp`, which recently added a `n_gpu_layers` argument. It would be great to have it in the wrapper.

Current workaround:
```
llm = LlamaCpp(...)
state = llm.client.__getstate__()
state["n_gpu_layers"] = n_gpu_layers
llm.client.__setstate__(state)
```

### Motivation

-

### Your contribution

-