Fix the link to cloud training in doc

Update default hyper-params of scorer in python scripts

Decouple data provider and model configuration.

fix a comment in audio_featurizer.py

Simplify parallel part for data processing and fix abnormal exit.

Retune hyper-parameters and update benchmark results for English models due to #50

Correct the error rate's computation for multiple sentences

Add infer & test scripts for baidu_en8k model

Expose edit distance for error_rate.py

Add library boost to the dependency

Fix doc for Docker

Update the setup in readme

Adapt demo_server.py to support padding removing.

Update libri model.

Add script for VoxForge data preparation.

Remove doc deploy in travis-ci