[TODO] Word tokenizition

[wannaphong]

Today, Thai have many word tokenizition model and It's hard to train own model, so I think we can add easy train model method.

Preprocessing

Get pandas.DataFrame to train/val/test model

| Domain | type(train/test/val) | text |
| news | train | ผม|กำลัง|อยู่|ที่|ทำเนียนัฐบาล |

and train/va/test by each domain.

Model

Train with

• Deepcut

and more...

Inference

Get onnx to running model.