clone from https://huggingface.co/datasets/lj_speech

and support set data_files


# use

```python

dataset = load_dataset('SeanSleat/lj_speech',data_files='/path/to/LJSpeech-1.1.tar.bz2')