Train a handwriting English model problem

[Robert Zhou]

I use Tesstrain instruction to train a handwriting model from starting tessdata English model. The training was finished without errors. I produced two traineddata files, one under data dir, another small size one under the model name dir. I test two model files with pytesseract. The one under data dir works like starting English model, cannot extract handwriting text. The small size one under model dir cannot be loaded by pytesseract. 

do you have any ideas what's the problem, how to get a working trained model file?