[Examples - DreamBooth] Add an option to tune the text encoder with LoRA 

We recently added support for fine-tuning the text encoder with LoRA here: https://github.com/huggingface/diffusers/pull/2918.

Next up, we need to modify our [DreamBooth training example](https://github.com/huggingface/diffusers/tree/main/examples/dreambooth) to incorporate this support. 