Created by: daming-lu

The right way is: fluid.io.save_inference_model

In model: https://github.com/PaddlePaddle/models/blob/develop/fluid/DeepASR/train.py#L278 https://github.com/PaddlePaddle/models/blob/develop/fluid/DeepQNetwork/train.py#L142 https://github.com/PaddlePaddle/models/blob/develop/fluid/text_classification/train.py#L80 https://github.com/PaddlePaddle/models/blame/develop/fluid/text_classification/train.py#L80

In paddle: https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/fluid/tests/book/test_recognize_digits.py#L140 https://github.com/PaddlePaddle/Paddle/blob/develop/python/paddle/fluid/tests/book/test_recommender_system.py#L213 https://github.com/PaddlePaddle/Paddle/blame/develop/python/paddle/fluid/tests/book/test_image_classification.py#L174

But in book: it uses trainer.save_params

https://github.com/PaddlePaddle/book/blame/develop/05.recommender_system/train.py#L184 https://github.com/PaddlePaddle/book/blob/develop/06.understand_sentiment/train_conv.py#L106 https://github.com/PaddlePaddle/book/blob/develop/04.word2vec/train.py#L113