1. Refine data_utils/data.py, reuse process_utterance function.
2. Modified README.
3. Implement uploading data in cloud/upload_data.py
4. Merge branch 'develop' of https://github.com/PaddlePaddle/models into ds2_pcloud

Add vocabulary dictionary building script

Unify encoding to 'utf-8' and optimize error rate calculation.

Make type of error rate optional.

Add NoisePerturbAugmentor and CHiME3 noise data preparation.

Accelerate mfcc computation for DS2.

Reset default multi-thread/process number to half of cpu count() for speedup.

Add a realtime ASR demo (both server and client) for DS2 users to try with own voice.

Simplify train.py, evaluate.py, infer.py and tune.py by adding DeepSpeech2Model class for DS2.

Add the requirement for cuDNN version in README

change the wget method in run.sh of deep_speech2

make kenlm install more robust

update installation of libsndfile, add unitest of soundfile