* mv dataio.py to s2t.io.speechbrain.dataio

mv dataio.py to paddlespeech.s2t.io.speechbrain.dataio

* remove transformers import.

* change optimizer same with released model

* add paddlenlp version in RESULT.md.

* fix run.sh

* fix data.sh step_num.

* add adadelta optimizer config.

* fix wav2vec2 test_wav.sh run error.

* add tokenizer config.