Simplifying data reading system

v1.1

• Support resampling in DumpReader

VCTK, Multiple readers, WavDataset supports

v0.5.1

• remove dump
• fix type hinting
• LibriSpeech support

v0.5.2

• dumped reader

Fix dumped dataset relative errors

v0.4

dump: Add chunksize

Make numpy major

Updates

1. LibriTTS supports
2. Multispeaker Wrapper
3. Dumped dataset supports
4. Add report level on text normalizer

Numpy-based speechset

numpy-based speechset for pytorch and jax supports

TFRecord support

Support tfrecord for preprocessing data

Initial release

v0.1

readme: Update readme